WO2011074529A1 - Time synchronization system, slave node, time synchronization method, and program for time synchronization - Google Patents

Abstract

A system is provided with a master node and a slave node which mutually communicate, and the time at the slave node is synchronized with the time at the master node. The slave node is provided with a transmission unit which transmits a control message to the master node; a reception unit which receives the control message from the master node; a delay measurement unit which measures a delay amount indicating a queueing delay which is received in a communication path by the control message transmitted from the master node to the slave node; a propagation delay measurement unit which measures a propagation delay amount between the master node and the slave node; and a time synchronization control unit which calculates a difference between the time at the slave node and the time at the master node using the delay amount measured by the delay measurement unit and the propagation delay amount measured by the propagation delay measurement unit, and synchronizes the time at the slave node with the time at the master node.

Description

Time synchronization system, slave node, time synchronization method, and time synchronization program

The present invention relates to a technique for synchronizing time between nodes that perform communication.
This application claims priority based on Japanese Patent Application No. 2009-287804 for which it applied on December 18, 2009, and uses the content here.

Conventionally, highly accurate time synchronization is required for mobile backhaul, factory automation, home AV LAN, and the like. Against this background, standardization of technology for performing time synchronization is underway. For example, there is IEEE 1588 as a technique for performing time synchronization with accuracy of less than microseconds in a packet network (see Non-Patent Document 1).

In the technology specified in IEEE1588, time stamp information is exchanged by exchanging messages between master / slave nodes. The slave node calculates a time lag (Offset) of the slave node with respect to the master node from the message transmission / reception times at the master and the slave node. Then, the slave node corrects the time of the slave node based on this Offset, and synchronizes the time of the slave node with the master node.

IEEE 1588 assumes that the transmission delay (MS_Delay) from the master node to the slave node is equal to the transmission delay (SM_Delay) from the slave node to the master node in order to obtain Offset. However, in a packet network, there is a high possibility that MS_Delay and SM_Delay have different values due to queuing delays in relay nodes such as routers and switches. For this reason, the assumption that MS_Delay and SM_Delay are equal causes an error in Offset, which degrades the time synchronization accuracy.

In order to solve such problems, IEEE 1588 version 2 defines the Transparent Clock (TC) function. In the following description, IEEE 1588 using this TC function is referred to as IEEE 1588v2 w / TC.

IEEE P1588TM / D1, "Draft Standard for Precision, Clock, Synchronization, Protocol, for Networked Measurement, and Control Systems," June, 2007.

However, in IEEE1588v2 w / TC, in order to achieve high accuracy, more relay nodes (ultimately all relay nodes) need to have a TC function. On the other hand, most existing nodes already placed in the network do not have the TC function. Therefore, in order to realize IEEE1588v2 w / TC, it is necessary to replace the relay node and add functions. Therefore, there is a problem that the realization of IEEE1588v2 w / TC is very difficult from the viewpoint of cost and labor.
In view of the above circumstances, an object of the present invention is to provide a technique for realizing highly accurate time synchronization with reduced costs.

The present invention is a time synchronization system that includes a master node and a slave node that communicate with each other, and synchronizes the time at the slave node with the time at the master node, wherein the master node transmits a control message to the slave node. And a receiver that receives a control message from the slave node, wherein the slave node transmits a control message to the master node, and a receiver that receives the control message from the master node A delay measurement unit that measures a delay amount representing a queuing delay received in a communication path by a control message transmitted from the master node to the slave node, and a propagation delay amount between the master node and the slave node Propagation delay measuring unit for measuring the delay and the delay measurement Using the delay amount measured by the propagation delay measurement unit and the propagation delay amount measured by the propagation delay measurement unit, the difference between the time at the slave node and the time at the master node is calculated, and the time at the slave node is A time synchronization system including a time synchronization control unit that synchronizes with a time in a master node is provided.

The present invention is also a slave node that synchronizes time with a master node that includes a transmission unit that transmits a control message to a slave node and a reception unit that receives a control message from the slave node, A transmission unit that transmits a control message to a node, a reception unit that receives the control message from the master node, and a queuing delay that the control message transmitted from the master node to the slave node receives in a communication path A delay measuring unit for measuring a delay amount, a propagation delay measuring unit for measuring a propagation delay amount between the master node and the slave node, a delay amount measured by the delay measuring unit, and the propagation delay measuring unit And the time at the slave node and the mass It calculates a difference between the time at node provides a slave node and a time synchronization control unit for synchronizing the time at the master node times in the slave node.

The present invention is also a time synchronization method performed by a time synchronization system that includes a master node and a slave node that communicate with each other, and that synchronizes the time at the slave node with the time at the master node, wherein the master node is connected to the slave node. A transmitting step of transmitting a control message, a receiving step of the master node receiving a control message from the slave node, a transmitting step of the slave node transmitting a control message to the master node, and the slave node A reception step of receiving the control message from the master node, and a delay in which the slave node measures a delay amount representing a queuing delay received in a communication path by the control message transmitted from the master node to the slave node. Total A propagation delay measuring step in which the slave node measures a propagation delay amount between the master node and the slave node; a delay amount measured by the slave node in the delay measuring step; and the propagation Time synchronization control for calculating the difference between the time at the slave node and the time at the master node using the propagation delay amount measured at the delay measurement step, and synchronizing the time at the slave node with the time at the master node And a time synchronization method comprising the steps.

The present invention also includes a master node and a slave node that communicate with each other, and corresponds to a first computer corresponding to the master node and a slave node as a time synchronization system that synchronizes the time in the slave node with the time in the master node. A time synchronization program for operating a second computer, a transmitting step for transmitting a control message to the slave node to the first computer, and a receiving step for receiving a control message from the slave node To the second computer, a transmission step of transmitting a control message to the master node, a reception step of receiving the control message from the master node, and from the master node to the slave node Send A delay measuring step for measuring a delay amount representing a queuing delay received by the control message in the communication path, a propagation delay measuring step for measuring a propagation delay amount between the master node and the slave node, and the delay measurement. The difference between the time at the slave node and the time at the master node is calculated using the delay amount measured at the step and the propagation delay amount measured at the propagation delay measurement step, and the time at the slave node is calculated. There is provided a time synchronization program for executing a time synchronization control step for synchronizing with a time in the master node.

The present invention makes it possible to realize highly accurate time synchronization with reduced costs.

FIG. 3 is a sequence diagram showing a communication sequence by an IEEE 1588 time synchronization algorithm. It is a figure showing the outline of the time synchronization algorithm of IEEE1588. It is a figure showing the outline of the time synchronization algorithm of IEEE1588v21w / TC. 1 is a system configuration diagram of a communication system. It is a functional block diagram showing the functional structure of a master node. It is a functional block diagram showing the functional structure of a slave node. It is a functional block diagram showing the detailed functional structure of a delay measurement part. It is the schematic showing the outline | summary of operation | movement of a packet counter. It is a functional block diagram showing the detailed functional structure of a time synchronous control part. It is a flowchart showing the flow of the master 1st process by a master node. It is a flowchart showing the flow of the initial mode slave 1st process by a slave node. It is a flowchart showing the flow of the master 2nd process by a master node. It is a flowchart showing the flow of the initial mode slave 2nd process by a slave node. It is a flowchart showing the flow of the normal mode slave 1st process by a slave node. It is a flowchart showing the flow of the normal mode slave 2nd process by a slave node. It is a sequence diagram showing the specific example of the time synchronization process by the communication system in an initial mode. Similarly, it is a sequence diagram showing a specific example of time synchronization processing by the communication system in the initial mode. Similarly, it is a sequence diagram showing a specific example of time synchronization processing by the communication system in the initial mode. It is a sequence diagram showing the specific example of the time synchronous process by the communication system in normal mode. It is a figure showing the uniform distribution of the delay amount measured in a communication system. It is a figure showing the Poisson distribution of the delay amount measured in a communication system. It is a figure showing the outline of a process of a delay measurement part at the time of comprising using a packet buffer.

[IEEE1588]
First, the IEEE 1588 time synchronization algorithm will be described. FIG. 1 is a sequence diagram showing a communication sequence according to the IEEE 1588 time synchronization algorithm.
In FIG. 1, the master node 100 and the slave node 200 perform bidirectional communication, and the slave node 200 periodically synchronizes the time with the master node 100.

The master node 100 periodically transmits a Sync message to the slave node 200 (step S100).
The master node 100 records the transmission time (hereinafter referred to as “Sync transmission time”) Tm (0) of this Sync message (step S101).
Next, the master node 100 transmits a Follow_up message to the slave node 200 (step S102). At this time, the master node 100 stores the Sync transmission time Tm (0) in the Follow_up message.

When the slave node 200 receives the Sync message, the reception time of the Sync message (hereinafter referred to as “Sync reception time”) Ts (0) is recorded using this reception process as a trigger (step S103).
Next, the slave node 200 receives the Follow_up message, extracts and records the Sync transmission time Tm (0) stored in the Follow_up message (step S104).
Next, the slave node 200 transmits a Delay_Request message to the master node 100 (step S105).
Then, the slave node 200 records the transmission time of this Delay_Request message (hereinafter referred to as “Delay transmission time”) Ts (1) (step S106).

When receiving the Delay_Request message, the master node 100 records the reception time of the Delay_Request message (hereinafter referred to as “Delay reception time”) Tm (1) using this reception process as a trigger (step S107).
Next, the master node 100 transmits a Delay_Response message to the slave node 200 (step S108). At this time, the master node 100 stores the Delay reception time Tm (1) in the Delay_Response message.

When receiving the Delay_Response message, the slave node 200 extracts and records the Delay reception time Tm (1) stored in the Delay_Response message (step S109).
Based on the Sync transmission time Tm (0) and the Sync reception time Ts (0), the slave node 200 calculates the time at the master node 100 (hereinafter referred to as “master time”) and the time at the slave node 200 from the following Expression 1. The difference MS_Diff from (hereinafter referred to as “slave time”) is calculated.
MS_Diff = Ts (0)-Tm (0) = MS_Delay + Offset ・ ・ ・ Equation 1

In addition, the slave node 200 obtains a difference between the slave time and the master time from the following formula 2 based on the Delay transmission time Ts (1) and the Delay reception time Tm (1).
SM_Diff = Tm (1)-Ts (1) = SM_Delay-Offset ・ ・ ・ Equation 2

Here, MS_Delay represents a transmission delay from the master node 100 to the slave node 200, SM_Delay represents a transmission delay from the slave node 200 to the master node 100, and Offset represents a time offset (advance) of the slave node 200 with respect to the master node 100. Represents.
The transmission delays MS_Delay and SM_Delay are composed of a propagation delay between the master node 100 and the slave node 200 and a queuing delay that occurs at a relay node on the network between the master node 100 and the slave node 200.

As described above, two expressions, Expression 1 and Expression 2, are obtained with respect to Offset, which is a time lag of the slave node 200 with respect to the master node 100. However, these two equations include unknown parameters MS_Delay and SM_Delay in addition to Offset. Therefore, since there are only two equations for the three unknown parameters, Offset cannot be calculated.
Therefore, in IEEE1588, assuming that the transmission delay MS_Delay from the master node 100 to the slave node 200 is equal to the transmission delay SM_Delay from the slave node 200 to the master node 100, and that both values are delays, 1 and 2 are transformed into the following equations 3 and 4.

MS_Diff = Delay + Offset ・ ・ ・ Equation 3
SM_Diff = Delay-Offset ・ ・ ・ Equation 4
By solving the simultaneous equations of Equation 3 and Equation 4, the following Equation 5 is derived.
Offset = (MS_Diff-SM_Diff) / 2 ... Formula 5

The slave node 200 calculates Offset based on Equation 5, and corrects the slave time based on Offset to synchronize the slave time with the master time. The above is the time synchronization algorithm defined in IEEE1588.

[IEEE1588v2 w / TC]
Next, the IEEE 1588v2 w / TC time synchronization algorithm will be described. FIG. 2A is a diagram showing an outline of the time synchronization algorithm of IEEE1588. FIG. 2B is a diagram showing an outline of the IEEE 1588v2 w / TC time synchronization algorithm. D1 to D6 in FIGS. 2A and 2B represent queuing delays in transmissions in the directions indicated by arrows in the drawings, which occur in the relay nodes Re1 to Re3, respectively.

In IEEE1588v2 w / TC, each relay node Re1 to Re3 has a TC function. The TC function is a function that measures the stay time in the node of the packet of the control message (IEEE 1588 message), writes the time in a predetermined field of the control packet, and cumulatively adds it. Note that the IEEE 1588 message is specifically a Sync message and a Delay_Request message.
In IEEE1588v2 w / TC, the stay time at the relay nodes Re1 to Re3 is cumulatively added to the message every time the control packet passes through the relay nodes Re1 to Re3 by the TC function. Therefore, the slave node 200 can accurately obtain the sum of the queuing delays that have occurred in the relay nodes Re1 to Re3 in the transmission from the master node 100 to the slave node 200. Similarly, the master node 100 can accurately obtain the sum of the queuing delays that have occurred in the relay nodes Re1 to Re3 in the transmission from the slave node 200 to the master node 100.

The total queuing delay and propagation delay in transmission from the master node 100 to the slave node 200 are MS_Q and MS_P, respectively, and the total queuing delay and propagation delay in transmission from the slave node 200 to the master node 100 are SM_Q, respectively. , SM_P, Equation 1 and Equation 2 described above can be transformed into Equation 6 and Equation 7 below.
MS_Diff = MS_P + MS_Q + Offset ・ ・ ・ Formula 6
SM_Diff = SM_P + SM_Q-Offset ・ ・ ・ Equation 7

Here, when the message transmission path of the transmission from the master node 100 to the slave node 200 and the transmission from the slave node 200 to the master node 100 are equal in both directions, MS_P = SM_P = Propagation_Delay. In this case, Equation 6 and Equation 7 can be transformed as Equation 8 and Equation 9 below.
MS_Diff = Propagation_Delay + MS_Q + Offset ・ ・ ・ Formula 8
SM_Diff = Propagation_Delay + SM_Q-Offset ・ ・ ・ Equation 9

Then, from Expression 8 and Expression 9, the following Expression 10 can be obtained as an expression for calculating Offset.
Offset = {(MS_Diff-SM_Diff)-(MS_Q-SM_Q)} / 2 Equation 10

As shown in FIG. 2A, IEEE1588 (hereinafter also referred to as “Pure IEEE1588”), which is not IEEE1588v2 w / TC, assumes that the sum of the queuing delays occurring at each relay node Re1 to Re3 is equal in both directions. It was. That is, the total queuing delay in transmission from the master node 100 to the slave node 200 (D1 + D2 + D3) and the total queuing delay in transmission from the slave node 200 to the master node 100 (D4 + D5 + D6) Was assumed to be equal. However, since they are not actually equal, the error has caused the deterioration of synchronization accuracy.

In contrast, IEEE 1588v2 w / TC measures the total queuing delay at each relay node Re1 to Re3 by the TC function implemented on each relay node Re1 to Re3. Then, the slave node 200 accurately acquires the total value (D1 + D2 + D3) of the queuing delay in the transmission from the master node 100 to the slave node 200. Further, the master node 100 accurately acquires the total value (D4 + D5 + D6) of the queuing delay in the transmission from the slave node 200 to the master node 100. With this operation, IEEE 1588v2 w / TC enables highly accurate time synchronization. The above is the time synchronization algorithm defined in IEEE1588v2 w / TC.

[System configuration and functional blocks]
FIG. 3 is a system configuration diagram of the communication system 1. The communication system 1 includes a master node 300, slave nodes 400 (400a to 400c), and a packet network PN. Hereinafter, details of the configurations of the master node 300 and the slave node 400 will be described.

[Master node]
FIG. 4 is a functional block diagram showing the functional configuration of the master node 300.
The master node 300 includes a clock generation unit 301, a master clock unit 302, a packet generation unit 303, a packet transmission unit 304, and a packet reception unit 305. The master node 300 includes, for example, a CPU (Central Processing Unit) connected via a bus, a memory, an auxiliary storage device, a communication interface, and the like, and is configured as a device including each of the above functional units by executing a time synchronization program. Also good.

The clock generation unit 301 generates a reference clock for the master node 300. Specifically, the clock generation unit 301 determines a time width of 1 second in the master node 300. Note that the clock generation unit 301 may exist outside the master node 300. In this case, the master node 300 is configured to reliably acquire a clock from the clock generation unit 301 installed outside.

The master clock unit 302 determines the time (master time) of the master node 300 according to the reference clock generated by the clock generation unit 301. Specifically, the master clock unit 302 determines what hours, minutes, and seconds in the master node 300.

The packet generation unit 303 generates an IEEE 1588 message and sends it to the packet transmission unit 304. The IEEE 1588 messages are specifically the above-mentioned Sync message, Follow_up message, and Delay_Response message.
The packet generation unit 303 periodically transmits a Sync message to the slave node 400 via the packet transmission unit 304, and simultaneously records the Sync transmission time Tm (0) with reference to the master clock unit 302.
Further, after transmitting the Sync message, the packet generation unit 303 generates a Follow_up message storing the Sync transmission time Tm (0), and transmits it to the slave node 400 via the packet transmission unit 304.
Further, the packet generation unit 303 generates a Delay_Response message storing the reception time (Delay reception time) Tm (1) of the Delay_Request message received by the packet reception unit 305, and sends it to the slave node 400 via the packet transmission unit 304. Send.

Also, the Delay_Response message is a reply to the Delay_Request message transmitted from the slave nodes 400a to 400c. Therefore, the packet generation unit 303 manages the Delay reception time Tm (1) stored in the Delay_Response message for each slave node 400a to 400c that is the source of the Delay_Request message.

The packet transmission unit 304 transmits the Sync message, Follow_up message, and Delay_Response message received from the packet generation unit 303 to the slave nodes 400a to 400c via the packet network PN.
Note that the packet transmission unit 304 broadcasts a Sync message and Follow_up message to the slave nodes 400a to 400c.
The packet transmission unit 304 unicasts the Delay_Response message to the slave nodes 400a to 400c that are the transmission sources of the Delay_Request messages.

The packet receiving unit 305 receives a Delay_Request message sent from the slave nodes 400a to 400c via the packet network PN. Then, the packet receiver 305 transfers the received Delay_Request message to the packet generator 303.

[Slave node]
FIG. 5 is a functional block diagram illustrating a functional configuration of the slave node 400. Each of the slave nodes 400a to 400c has the same configuration as that of the slave node 400.
The slave node 400 includes a packet reception unit 401, a PLL unit 402, a delay measurement unit 403, a slave clock unit 404, a time synchronization control unit 405, and a packet transmission unit 406.
The slave node 400 may include, for example, a CPU, a memory, an auxiliary storage device, a communication interface, and the like connected by a bus, and may be configured as a device including the above-described functional units by executing a time synchronization program.

The packet receiving unit 401 receives the Sync message, Follow_up message, and Delay_Request message transmitted from the master node 300 via the packet network PN, and transfers them to the time synchronization control unit 405. The packet reception unit 401 transfers only the Sync message to the PLL unit 402 and the delay measurement unit 403.

The PLL unit 402 includes a phase comparator 407, an LPF 408, a PI controller 409, a VCO 410, and a counter 411. However, such a configuration of the PLL unit 402 is merely an example. That is, the PLL unit 402 calculates the difference between the time generated from its own clock and the time received from the master node 300, and any other configuration can adjust its own clock based on the difference. Such a configuration may be adopted.

The phase comparator 407 calculates a difference signal between the reception time stamp stored in the Sync message packet received from the packet reception unit 401 and the time stamp generated by the counter 411. Then, the phase comparator 407 outputs a differential signal representing the calculation result to the LPF 408.

The LPF 408 leveles the differential signal output by the phase comparator 407, suppresses jitter and noise, and outputs the result to the PI controller 409.
The PI controller 409 generates a control signal such that the leveled difference signal finally becomes zero, and outputs the control signal to the VCO 410.

The VCO 410 generates a clock having a frequency determined by the control signal output from the PI controller 409 and outputs the clock to the counter 411. In addition, the VCO 410 decreases the counter value of the packet counter of the delay measurement unit 403 according to the generated frequency clock.
The counter 411 generates a time stamp based on the clock generated by the VCO 410 and transfers the time stamp to the phase comparator 407.

With the operation of the PLL unit 402 described above, the clock of the slave node 400 generated by the VCO 410 is synchronized with the clock of the master node 300.
The delay measurement unit 403 includes a packet counter. The delay measuring unit 403 calculates the delay amount MS_Q of the incoming Sync message by monitoring the increase / decrease state of the counter value of the packet counter. Then, the delay measurement unit 403 notifies the time synchronization control unit 405 of the calculated delay amount MS_Q.
Note that the delay amount MS_Q represents the delay amount of the queuing delay received while the Sync message is transferred from the master node 300 to the slave node 400 as described above.

FIG. 6 is a functional block diagram illustrating a detailed functional configuration of the delay measurement unit 403.
The delay measurement unit 403 includes a packet counter 501, a counter maximum value monitor unit 502, an arrival counter value monitor unit 503, and a delay calculation unit 504.
The counter maximum value monitoring unit 502 and the arrival time counter value monitoring unit 503 monitor the counter value of the packet counter 501 and notify the delay calculation unit 504 of the results to be described later.
The delay calculation unit 504 calculates the delay amount MS_Q using information notified from the counter maximum value monitor unit 502 and the arrival time counter value monitor unit 503. Then, the delay calculation unit 504 notifies the time synchronization control unit 405 of the calculated delay amount MS_Q.

Each time the packet counter 501 receives a Sync message from the packet receiver 401, the packet counter 501 increases the counter value by a predetermined value. Further, the packet counter 501 decreases the counter value in accordance with the frequency of the clock output from the VCO 410.
The reason why the counter value of the packet counter 501 decreases according to the clock frequency of the VCO 410 is that the Sync message packet received by the packet receiving unit 401 is read from the buffer of the received packet according to the clock frequency of the VCO 410.

The counter maximum value monitor unit 502 detects the maximum value of the counter value of the packet counter 501. Specifically, the counter maximum value monitoring unit 502 detects the maximum value within a monitoring period of a predetermined time (for example, 10 seconds). Hereinafter, the maximum value of the counter in the monitoring period i is expressed as a counter maximum value P (i).
When the monitoring period i ends, the counter maximum value monitoring unit 502 notifies the delay calculating unit 504 of the counter maximum value P (i) that is the detection result.

The arrival time counter value monitoring unit 503 detects the arrival time counter value C (i, n) when the n-th Delay_Request message packet arrives in the monitoring period i. The arrival time counter value monitoring unit 503 detects the counter value C (i, n) every time the Delay_Request message packet arrives, and notifies the delay calculation unit 504 of the detection result.

The delay calculation unit 504 uses the counter value C (i, n) and the counter maximum value P (i) each time the counter value C (i, n) is received from the arrival counter value monitoring unit 503, and the following equation is obtained. 11 is used to calculate the delay amount MS_Q.
MS_Q = P (i-1)-C (i, n) Equation 11
Here, P (i-1) is the maximum counter value in the previous monitoring period.
The delay calculation unit 504 notifies the time synchronization control unit 405 of the delay amount MS_Q every time the delay amount MS_Q is calculated.

FIG. 7 is a schematic diagram showing an outline of the operation of the packet counter 501.
In FIG. 7, the vertical axis represents the counter value, and represents a high counter value from the bottom to the top. The horizontal axis represents time and progresses from left to right. A plurality of broken lines arranged at equal intervals extending in the vertical direction indicate the timing at which a Sync message packet is received when no queuing delay occurs. An arrow pointing upward indicates the timing at which a Sync message packet is actually received.
When receiving the packet of the Sync message, the packet counter 501 increases the counter value by a fixed value (CV in FIG. 7). Thereafter, since the received packet is periodically read out by the processing unit at the subsequent stage, the counter value decreases at a constant pace.
When the vertically extending broken line and the upward arrow coincide with each other on the time axis, that is, when a Sync message packet is received without causing a queuing delay, the counter value takes the maximum value Cmax. In this case, in the packet counter 501, the total number of bits received matches the total number of bits read periodically according to the clock frequency of the VCO 410.
On the other hand, when a queuing delay occurs, the counter value is smaller than Cmax. At this time, the difference between the maximum value Cmax and the counter value is equivalent to the delay amount of the generated queuing delay.

Referring back to FIG. 5, the description of the configuration of the slave node 400 will be continued. The slave clock unit 404 determines the time of the slave node 400 (slave time) according to the clock generated by the VCO 410.

The time synchronization control unit 405 includes transmission / reception time information (Sync transmission time Tm (0), Sync reception time Ts (0), Delay transmission time Ts (1), Delay reception time Tm (1)) of the Sync message and Delay_Request message, Using the delay amount MS_Q, an offset which is a time lag of the slave node 400 with respect to the master node 300 is obtained. Then, the time synchronization control unit 405 corrects the time of the slave clock unit 404 (slave time). Also, the time synchronization control unit 405 generates a Delay_Request message and sends it to the packet transmission unit 406.
The packet transmission unit 406 transmits the Delay_Request message received from the time synchronization control unit 405 to the master node 300 via the packet network PN.

FIG. 8 is a functional block diagram illustrating a detailed functional configuration of the time synchronization control unit 405.
The time synchronization control unit 405 includes a packet analysis unit 701, a packet generation unit 702, an RTT calculation unit 703, a propagation delay monitor unit 704, an offset calculation unit 705, and a time adjustment unit 706.
The packet analysis unit 701 receives a Sync message, Follow_up message, and Delay_Response message from the packet reception unit 401. When receiving the Sync message, the packet analysis unit 701 refers to the slave clock unit 404 and acquires the Sync reception time Ts (0).
Then, the packet analysis unit 701 notifies the RTT calculation unit 703 and the offset calculation unit 705 of the acquired Sync reception time Ts (0).

Further, when receiving the Follow_up message, the packet analysis unit 701 extracts the Sync transmission time Tm (0) stored in the Follow_up message. Then, the packet analysis unit 701 notifies the extracted Sync transmission time Tm (0) to the RTT calculation unit 703 and the offset calculation unit 705, and sends a generation trigger for a Delay_Request message to the packet generation unit 702.

Further, when receiving the Delay_Response message, the packet analysis unit 701 extracts the Delay reception time Tm (1) stored in the Delay_Response message and notifies the RTT calculation unit 703 of it.

The packet generation unit 702 generates a Delay_Request message. Specifically, upon receiving a Delay_Request message generation trigger from the packet analysis unit 701, the packet generation unit 702 generates a Delay_Request message. The packet generation unit 702 transfers the generated Delay_Request message to the packet transmission unit 406 and refers to the slave clock unit 404 to acquire the Delay transmission time Ts (1). Then, the packet generator 702 notifies the RTT calculator 703 of the acquired Delay transmission time Ts (1).

The RTT calculation unit 703 calculates an RTT value (round trip (delay) time), and notifies the propagation delay monitor unit 704 of the calculated RTT value. Specifically, the RTT calculation unit 703 receives the Sync transmission time Tm (0), the Sync reception time Ts (0), the Delay transmission time Ts (1), and the Delay reception Tm received from the packet analysis unit 701 and the packet generation unit 702. Calculate the RTT value using (1). Note that the RTT calculation unit 703 calculates an RTT value in an initial mode process to be described later.

The propagation delay monitor unit 704 detects the propagation delay amount P and notifies the offset calculation unit 705 of the detected propagation delay amount P. Specifically, the propagation delay monitor unit 704 transmits the RTT value in each series of IEEE1588 messages (a set of Sync message, Follow_up message, Delay_Request message, and Delay Response message) transmission / reception (message exchange) and the message exchange. The difference from the delay amount MS_Q of the Sync message in is calculated as a propagation delay amount candidate X.
The propagation delay monitor unit 704 repeatedly calculates the propagation delay amount candidate X for each message exchange that occurs within a predetermined period.
Then, the propagation delay monitoring unit 704 determines the half of the minimum value among the propagation delay amount candidates X obtained within a predetermined period as the propagation delay amount P. Note that the propagation delay monitor unit 704 also detects the propagation delay amount P in the initial mode processing described later, similarly to the RTT calculation unit 703 described above.

The offset calculation unit 705 calculates Offset and notifies the time adjustment unit 706 of the calculated Offset. Specifically, offset calculation section 705 uses Sync transmission time Tm (0), Sync reception time Ts (0) received from packet analysis section 701, and propagation delay amount P received from propagation delay monitoring section 704, Calculate Offset. The offset calculation unit 705 calculates Offset in the normal mode process described later.

The time adjustment unit 706 adjusts the slave time of the slave clock unit 404 using the offset notified from the offset calculation unit 705. With this processing of the time adjustment unit 706, the slave time of the slave clock unit 404 is synchronized with the master time of the master clock unit 302 of the master node 300.

[Operation]
Next, the process flow of each device of the communication system 1 will be described.
The processing of the communication system 1 includes processing in the initial mode and processing in the normal mode. The communication system 1 first measures the propagation delay amount P by repeatedly operating for a predetermined period in the initial mode. Thereafter, the communication system 1 repeatedly operates in the normal mode to obtain an offset using the propagation delay amount P and synchronizes the slave time with the master time.
In the initial mode, the communication system 1 performs a master first process by the master node 300, an initial mode slave first process by the slave node 400, a master second process by the master node 300, and an initial mode slave second process by the slave node 400. Run each in this order.
In the normal mode, the communication system 1 performs a master first process by the master node 300, a normal mode slave first process by the slave node 400, a master second process by the master node 300, and a normal mode slave second process by the slave node 400. Run each in this order.

9 and 11 are flowcharts showing the flow of the master first process and the master second process by the master node 300, respectively.
10 and 12 are flowcharts showing the flow of the initial mode slave first process and the initial mode slave second process by the slave node 400, respectively.
FIGS. 13 and 14 are flowcharts showing the flow of the normal mode slave first process and the normal mode slave second process by the slave node 400, respectively.
The slave nodes 400a to 400c individually execute the processes shown in FIGS. 10, 12, 13 and 14.

First, the master first process by the master node 300 will be described with reference to FIG.
First, the master node 300 transmits a Sync message to the slave node 400 (step S201). Specifically, the packet generation unit 303 generates a Sync message, and the packet transmission unit 304 broadcasts the Sync message to the slave nodes 400 (400a to 400c).
Next, the master node 300 records the Sync transmission time Tm (0) (step S202). Specifically, the packet generation unit 303 refers to the master clock unit 302 to acquire and hold the Sync transmission time Tm (0).
Then, the master node 300 transmits a Follow_up message to the slave node 400 (step S203). Specifically, the packet generation unit 303 generates a Follow_up message storing the Sync transmission time Tm (0), and the packet transmission unit 304 broadcast-transmits the Follow_up message to the slave nodes 400 (400a to 400c).
The master node 300 periodically and repeatedly executes this master first process.

Next, the initial mode slave first processing by the slave node 400 will be described with reference to FIG.
First, the slave node 400 receives a Sync message from the master node 300 (step S301). Specifically, the packet reception unit 401 receives a Sync message from the master node 300 and transfers the received Sync message to the PLL unit 402, the delay measurement unit 403, and the time synchronization control unit 405.
In the operation after receiving the Sync message, the processes in steps S302 to S306 and the processes in steps S307 and S308 are executed in parallel.

First, the processing in steps S302 to S306 will be described.
First, the slave node 400 records the Sync reception time Ts (0) (step S302). Specifically, the packet analysis unit 701 of the time synchronization control unit 405 acquires the Sync reception time Ts (0) by referring to the slave clock unit 404 in response to the reception of the Sync message, and notifies the RTT calculation unit 703 of it. To do.

Next, the slave node 400 receives a Follow_up message from the master node 300 (step S303). Specifically, the packet reception unit 401 receives a Follow_up message from the master node 300 and transfers the received Follow_up message to the time synchronization control unit 405.
Next, the slave node 400 records the Sync transmission time Tm (0) (step S304). Specifically, the packet analysis unit 701 of the time synchronization control unit 405 extracts and acquires the Sync transmission time Tm (0) stored in the payload part of the Follow_up message in response to reception of the Follow_up message. Then, the packet analysis unit 701 notifies the RTT calculation unit 703 of the acquired Sync transmission time Tm (0).

Next, the slave node 400 transmits a Delay_Request message to the master node 300 (step S305). Specifically, the packet analysis unit 701 sends a Delay_Request message generation trigger to the packet generation unit 702. As a result, the packet generation unit 702 generates a Delay_Request message in response to the generation trigger, and transfers it to the packet transmission unit 406. Then, the packet transmission unit 406 transmits a Delay_Request message to the master node 300.
At this time, the slave node 400 records the Delay transmission time Ts (1) (step S306). Specifically, the packet generation unit 702 refers to the slave clock unit 404, acquires the Delay transmission time Ts (1), and notifies the RTT calculation unit 703 of it.

Next, the processing of steps S307 and S308 will be described.
First, the slave node 400 measures the delay amount MS_Q (step S307). Specifically, the counter maximum value monitor unit 502 and the arrival time counter value monitor unit 503 of the delay measurement unit 403 monitor the counter value of the packet counter 501, and each time the maximum counter value and each Sync message arrive in a predetermined interval period. Get the counter value. Then, the delay calculation unit 504 of the delay measurement unit 403 calculates the delay amount MS_Q based on both counter values.

Next, the slave node 400 records the measured delay amount MS_Q (step S308). Specifically, the delay calculation unit 504 notifies the calculated delay amount MS_Q to the propagation delay monitor unit 704 of the time synchronization control unit 405.

Next, the master second process by the master node 300 will be described with reference to FIG.
First, the master node 300 receives a Delay_Request message from the slave node 400 (step S401). Specifically, the packet reception unit 305 receives a Delay_Request message from the slave node 400 and transfers the received Delay_Request message to the packet generation unit 303.

Next, the master node 300 records the delay reception time Tm (1) (step S402). Specifically, the packet generation unit 303 refers to the master clock unit 302 in response to receiving the Delay_Request message, and acquires the time Tm (1) at the time of receiving the Delay_Request message. Then, the packet generation unit 303 records the Delay reception time Tm (1) of each Delay_Request message in association with the node information of the slave nodes 400a to 400c that is the transmission source of each Delay_Request message.

Next, the master node 300 transmits a Delay_Response message storing the Delay reception time Tm (1) to each of the slave nodes 400a to 400c (step S403).
Specifically, the packet generation unit 303 generates a Delay_Response message storing time information Tm (1) corresponding to each of the slave nodes 400a to 400c, and transfers it to the packet transmission unit 304. Then, the packet transmission unit 304 unicasts each Delay_Response message to the slave nodes 400a to 400c.

Next, the initial mode slave second processing by the slave node 400 will be described with reference to FIG.
First, the slave node 400 receives a Delay_Response message from the master node 300 (step S501). Specifically, the packet receiving unit 401 receives a Delay_Response message from the master node 300 and transfers the received Delay_Response message to the time synchronization control unit 405.

Next, the slave node 400 extracts and records the delay reception time Tm (1) from the received delay_response message (step S502). Specifically, the packet analysis unit 701 of the time synchronization control unit 405 extracts and acquires the Delay reception time Tm (1) stored in the payload part of the Delay_Response message in response to reception of the Delay_Response message. Then, the packet analysis unit 701 notifies the RTT calculation unit 703 of the acquired delay reception time Tm (1).

Next, the slave node 400 calculates an RTT value (step S503). Specifically, the RTT calculation unit 703 receives from the packet analysis unit 701 and the packet generation unit 702 the Sync transmission time Tm (0), the Sync reception time Ts (0), the Delay transmission time Ts (1), and the Delay reception time. Using Tm (1), the RTT value is calculated based on Equation 12 below.
RTT = Tm (1)-Tm (0)-(Ts (1)-Ts (0)) Equation 12
Thereafter, the RTT calculation unit 703 notifies the propagation delay monitor unit 704 of the calculated RTT value.

Next, the slave node 400 calculates a propagation delay amount candidate X for each sequence (a series of message exchanges) (step S504).
The slave node 400 repeatedly executes the calculation of the propagation delay amount candidate X for each message exchange within a predetermined period (step S505). Specifically, in one series of message exchanges, the propagation delay monitor unit 704 calculates one propagation delay amount candidate X based on the following Expression 13. Then, the propagation delay monitor unit 704 repeatedly calculates the propagation delay amount candidate X for each message exchange within a predetermined period.
X = RTT-MS_Q Equation 13

When the predetermined period has elapsed (step S505-YES), the slave node 400 determines the propagation delay amount P based on the propagation delay amount candidate X calculated within the predetermined period (step S506). Specifically, the propagation delay monitor unit 704 detects the minimum value of the propagation delay amount candidates X calculated within a predetermined period as the round trip propagation delay amount 2P. Then, the propagation delay monitor unit 704 determines the propagation delay amount P by halving the round trip propagation delay amount 2P.

The reason why the minimum value Xmin of the propagation delay amount candidate X becomes the round trip propagation delay amount 2P is as follows.
The RTT value includes a transmission delay from the master node 300 to the slave node 400 (propagation delay + queuing delay = P + MS_Q) and a transmission delay from the slave node 400 to the master node 300 (propagation delay + queuing delay = P +). Since it is the sum of SM_Q), Expression 14 is satisfied.
RTT = 2P + MS_Q + SM_Q Equation 14
Here, when MS_Q is moved to the left side, Expression 14 can be transformed into Expression 15.
RTT-MS_Q = 2P + SM_Q Equation 15

As is clear from Equation 15, the propagation delay amount candidate X is equal to 2P + SM_Q. Here, the round trip propagation delay amount 2P is a fixed value, and the queuing delay amount SM_Q is a variable value. The delay amount SM_Q may become zero depending on the network conditions.
Therefore, when a plurality of propagation delay amount candidates X are measured within a predetermined period and the minimum value (Xmin) is obtained as described above, the value is the propagation delay amount candidate X when the delay amount SM_Q = 0. The possibility is very high and can be approximated as such. The propagation delay amount candidate X in the case of SM_Q = 0 is X = 2P, and the one-way propagation delay amount P = Xmin / 2.

After calculating the propagation delay amount P, the propagation delay monitoring unit 704 notifies the offset calculation unit 705 of the calculated propagation delay amount P.
The predetermined period in step S505 is set to be shorter than the time for operating in the initial mode. That is, the process in the initial mode ends after the propagation delay amount P is calculated by the process in step S506 and notified to the offset calculation unit 705 after the elapse of a predetermined period in step S505.

Next, the normal mode slave first process by the slave node 400 will be described with reference to FIG.
First, the slave node 400 receives a Sync message from the master node 300. Specifically, the packet reception unit 401 receives a Sync message from the master node 300 and transfers the received Sync message to the PLL unit 402, the delay measurement unit 403, and the time synchronization control unit 405. In the operation after receiving the Sync message, the processes in steps S602 to S609 and the processes in steps S610 and S611 are executed in parallel.

First, the processing of steps S602 to S609 will be described. First, the slave node 400 records the Sync reception time Ts (0) (step S602). Specifically, the packet analysis unit 701 of the time synchronization control unit 405 acquires the Sync reception time Ts (0) by referring to the slave clock unit 404 in response to the reception of the Sync message, and notifies the offset calculation unit 705 To do.

Next, the slave node 400 receives a Follow_up message from the master node 300 (step S603). Specifically, the packet reception unit 401 receives a Follow_up message from the master node 300 and transfers the received Follow_up message to the time synchronization control unit 405.
Next, the slave node 400 records the Sync transmission time Tm (0) (step S604). Specifically, the packet analysis unit 701 of the time synchronization control unit 405 extracts and acquires the Sync transmission time Tm (0) stored in the payload part of the Follow_up message in response to reception of the Follow_up message. Then, the packet analysis unit 701 notifies the obtained Sync transmission time Tm (0) to the offset calculation unit 705. Thereafter, the packet analysis unit 701 notifies the packet generation unit 702 of a delay_request message generation trigger.

Next, the slave node 400 calculates a transmission delay MS_Diff (step S605). Specifically, the offset calculation unit 705 uses the Sync transmission time Tm (0) and the Sync reception time Ts (0) to perform transmission between the master node 300 and the slave node 400 based on the following Expression 16. Calculate and hold the delay MS_Diff.
MS_Diff = Ts (0)-Tm (0) Equation 16

Next, the slave node 400 calculates Offset (step S606). Specifically, the offset calculation unit 705 calculates Offset based on Expression 17 below using the transmission delay MS_Diff, the delay amount MS_Q, and the propagation delay amount P. Then, the offset calculation unit 705 notifies the time adjustment unit 706 of the calculated offset.

Offset = MS_Diff-P-MS_Q Equation 17
Equation 17 is obtained as follows.
Using MS_Diff, delay amount MS_Q, and propagation delay P measured in the initial mode slave second process, the following Expression 18 is established.
MS_Diff = P + MS_Q + Offset ・ ・ ・ Equation 18
Then, the above equation 17 can be obtained by transforming the equation 18.

Next, the slave node 400 adjusts the slave time based on Offset (step S607). Specifically, the time adjustment unit 706 delays the slave time by Offset when Offset is positive, and advances the slave time by Offset when Offset is negative. By adjusting the slave time, the slave time is synchronized with the master time of the master node 300.

Next, the slave node 400 transmits a Delay_Request message to the master node 300 (step S608). Specifically, the packet analysis unit 701 sends a Delay_Request message generation trigger to the packet generation unit 702. The packet generation unit 702 generates a Delay_Request message according to the generation trigger, and transfers it to the packet transmission unit 406. Then, the packet transmission unit 406 transmits a Delay_Request message to the master node 300.
At this time, the slave node 400 records the Delay transmission time Ts (1) (step S609). Specifically, the packet generation unit 702 refers to the slave clock unit 404, acquires the Delay transmission time Ts (1), and notifies the offset calculation unit 705 of it.

Next, the processing of steps S610 and S611 will be described.
First, the slave node 400 measures the delay amount MS_Q (step S610). Specifically, the counter maximum value monitor unit 502 and the arrival time counter value monitor unit 503 of the delay measurement unit 403 monitor the counter value of the packet counter 501, and each time the maximum counter value and each Sync message arrive in a predetermined interval period. Get the counter value. Then, the delay calculation unit 504 of the delay measurement unit 403 calculates the delay amount MS_Q based on both counter values.

Next, the slave node 400 records the measured delay amount MS_Q (step S611). Specifically, the delay calculation unit 504 notifies the calculated delay amount MS_Q to the offset calculation unit 705 of the time synchronization control unit 405. The notified delay amount MS_Q is used in the calculation of Offset in S606.

Next, the normal mode slave second processing by the slave node 400 will be described with reference to FIG.
First, the slave node 400 receives a Delay_Response message (step S701). Specifically, the packet reception unit 401 receives a Delay_Response message from the master node 300 and transfers the received Delay_Response message to the time synchronization control unit 405.

Next, the slave node 400 extracts the delay reception time Tm (1) (step S702).
Specifically, the packet analysis unit 701 of the time synchronization control unit 405 extracts and acquires the Delay reception time Tm (1) stored in the payload part of the Delay_Response message. Then, the packet analysis unit 701 notifies the obtained delay reception time Tm (1) to the offset calculation unit 705.

The operation has been described above.
In the initial mode, processing is executed in the order of master first processing, initial mode slave first processing, master second processing, and initial mode slave second processing.
In the normal mode, processing is executed in the order of master first processing, normal mode slave first processing, master second processing, and normal mode slave second processing.

The normal mode is executed to obtain an offset and adjust the slave time, but the process is completed in step S607 in FIG. That is, since the propagation delay P is calculated in the initial mode, the offset can be calculated only in the Sync reception time Tm (0) and the Sync transmission time Ts (0) in the normal mode. Therefore, in the normal mode, the Offset can be derived by exchanging only one-way Sync messages.
In the description of the normal mode described above, processing after the Delay_Request message (steps S608 and S609 in FIG. 13, FIGS. 11 and 14) has been described in consideration of consistency with the message exchange of the standard IEEE1588v2 w / TC. There is no need to perform this process.

15 to 17 are sequence diagrams showing specific examples of time synchronization processing by the communication system 1 in the initial mode. In the following description, an example in which three cycles of message exchange are performed as an initial mode will be described.
15 shows message exchange in the first cycle in the initial mode, FIG. 16 shows message exchange in the second cycle in the initial mode, and FIG. 17 shows message exchange in the third cycle in the initial mode.

15 to 17, the offset of the slave node 400 with respect to the master node 300 is “3”, and the propagation delay amount P between the master node 300 and the slave node 400 is “3”.
Further, the delay amount MS_Q of the queuing delay from the master node 300 to the slave node 400 is “2” in FIG. 15, “1” in FIG. 16, and “2” in FIG.
Further, the delay amount SM_Q of the queuing delay from the slave node 400 to the master node 300 is “2” in FIG. 15, “4” in FIG. 16, and “0” in FIG.
15 to 17, for convenience of explanation, the master time and the slave time are indicated by integer values instead of the notation of hour / minute / second.

In FIG. 15, the Sync transmission time Tm (0) in the first master process is “4”. In the initial mode slave first process, the Sync reception time Ts (0) is “12”, and the Delay transmission time Ts (1) is “16”. In the second master process, the delay reception time Tm (1) is “18”.
The delay amount MS_Q of the queuing delay from the master node 300 to the slave node 400 is measured as “2” by the delay measuring unit 403 in the initial mode slave first process. The RTT calculation unit 703 uses the above values to calculate the RTT value as follows.
RTT = (Tm (1) -Tm (0))-{(Ts (1)-Ts (0))} = 18-4-(16-12) = 14-4 = 10

Further, the propagation delay monitor unit 704 obtains a propagation delay amount candidate X as follows. Here, since the propagation delay amount candidate X obtained in FIG. 15 is a value in the first cycle of the initial mode, it is obtained as follows when represented as propagation delay amount candidate X (1).
X (1) = RTT-MS_Q = 10-2 = 8

In FIG. 16, the Sync transmission time Tm (0) in the first master process is “4”. In the initial mode slave first process, the Sync reception time Ts (0) is “11”, and the Delay transmission time Ts (1) is “15”. In the second master process, the delay reception time Tm (1) is “19”.
The delay amount MS_Q of the queuing delay from the master node 300 to the slave node 400 is measured as “1” by the delay measuring unit 403 in the initial mode slave first process. The RTT calculation unit 703 uses the above values to calculate the RTT value as follows.
RTT = (Tm (1)-Tm (0))-{(Ts (1)-Ts (0))} = 19-4-(15-11) = 15-4 = 11

Further, the propagation delay monitor unit 704 obtains a propagation delay amount candidate X as follows. Here, since the propagation delay amount candidate X obtained in FIG. 16 is a value in the second cycle of the initial mode, it is obtained as follows when represented as propagation delay amount candidate X (2).
X (2) = RTT-MS_Q = 11-1 = 10

In FIG. 17, the Sync transmission time Tm (0) in the first master process is “4”. In the initial mode slave first process, the Sync reception time Ts (0) is “12”, and the Delay transmission time Ts (1) is “15”. In the second master process, the delay reception time Tm (1) is “15”.
The delay amount MS_Q of the queuing delay from the master node 300 to the slave node 400 is measured as “2” by the delay measuring unit 403 in the initial mode slave first process. The RTT calculation unit 703 uses the above values to calculate the RTT value as follows.
RTT = (Tm (1)-Tm (0))-{(Ts (1)-Ts (0))} = 15-4-(15-12) = 11-3 = 8

Further, the propagation delay monitor unit 704 obtains a propagation delay amount candidate X as follows. Here, since the propagation delay amount candidate X obtained in FIG. 17 is a value in the third cycle of the initial mode, it is obtained as follows when represented as propagation delay amount candidate X (3).
X (3) = RTT-MS_Q = 8-2 = 6

By the above three measurements in the initial mode, X (1) = 8, X (2) = 10, X (3) = 6, and the minimum value of the propagation delay amount candidate X is “6”. As described above, when X = 2P + SM_Q, and when X is the minimum value, it is considered that SM_Q 考 え = 0. Therefore, from X = 2P, the propagation delay amount P is obtained as “3”. This indicates that the propagation delay “3” of the set condition is correctly obtained.

FIG. 18 is a sequence diagram illustrating a specific example of time synchronization processing by the communication system 1 in the normal mode. In FIG. 18, the offset of the slave node 400 with respect to the master node 300 is “3”, and the propagation delay amount P between the master node 300 and the slave node 400 is “3”, which is the same as in FIGS.
Further, the delay amount MS_Q of the queuing delay from the master node 300 to the slave node 400 is “2”, and the delay amount SM_Q of the queuing delay from the slave node 400 to the master node 300 is “3”.

In FIG. 18, the Sync transmission time Tm (0) in the master first process is “4”. In the first normal mode slave process, the Sync reception time Ts (0) is “12”. The delay amount MS_Q of the queuing delay from the master node 300 to the slave node 400 is measured as “2” by the delay measurement unit 403 in the normal mode slave first process.

The offset calculation unit 705 calculates Offset using the above values as follows.
Offset = MS_Diff-P-MS_Q
= (12-4)-3-2
= 3

In FIG. 18, since the set offset is “3”, it can be seen that the offset is accurately obtained from the transmission / reception of only the Sync message by the time synchronization processing of the communication system 1.

19A and 19B are diagrams illustrating probability distributions of delay amounts measured in the communication system 1. In the experiments in FIGS. 19A and 19B, a network emulator is connected instead of the packet network PN, a delay amount is added to the Sync message packet periodically transmitted from the master node 300, and the delay measurement unit in the slave node 400 At 403, the amount of delay was measured.
19A and 19B show the distribution of the delay amount added by the network emulator and the distribution of the delay amount measured by the slave node 400, respectively. In each figure, the horizontal axis represents the delay amount, and the vertical axis represents the probability.
In both cases of the uniform distribution (FIG. 19A) and the Poisson distribution (FIG. 19B), the distribution of the measured delay amount substantially coincides with the delay amount distribution added by the network emulator. Therefore, it is clear that the delay measurement unit 403 can correctly measure the delay amount MS_Q. Therefore, it is clear that the offset value is accurately calculated by the time synchronization process by the communication system 1, and the slave time can be accurately synchronized with the master time.

According to the communication system 1 described above, the delay measurement unit 403 of the slave node 400 measures the delay amount MS_Q of the queuing delay from the master node 300 to the slave node 400 based on the transmission / reception of the Sync message.
Further, the propagation delay monitor unit 704 of the slave node 400 measures the propagation delay amount P between the master node 300 and the slave node 400.
Then, the offset calculation unit 705 of the slave node 400 calculates Offset using the Sync transmission time Tm (0) and the Sync reception time Ts (0), the delay amount MS_Q, and the propagation delay amount P.
This eliminates the error caused by the actual difference in bidirectional queuing delay compared to the case of calculating Offset assuming that the bidirectional queuing delay is equal as in Pure IEEE1588. It becomes possible to realize time synchronization with higher accuracy.

Further, in the communication system 1, the delay amount of the queuing delay in each direction can be measured only by the master node 300 and the slave node 400. Therefore, it is not necessary to implement a special function such as the TC function in the relay node in the packet network PN, and it is possible to cope with the existing relay node.
In other words, compared to IEEE 1588v2 w / TC, which needs to implement the TC function for the relay node, there is no need to replace the relay node at all, the cost required for the introduction can be suppressed, and the introduction is easy There is.
In the communication system 1, in the normal mode, it is possible to realize time synchronization with high accuracy as described above only by transmitting and receiving a Sync message.

<Modification>
The delay measurement unit 403 may be configured using a packet buffer instead of the packet counter 501.
When receiving the packet of the Delay_Request message from the packet receiving unit 401, the packet buffer accumulates the received packet in the buffer. The packet buffer outputs the accumulated packets according to the frequency determined by the VCO 410.
FIG. 20 is a diagram illustrating an outline of processing of the delay measurement unit 403 when configured using a packet buffer.
In FIG. 20, the vertical axis represents the accumulation amount in the buffer, and represents a high accumulation amount from bottom to top. The horizontal axis represents time and progresses from left to right.
A plurality of broken lines arranged at equal intervals extending in the vertical direction indicate the timing at which a packet of a Delay_Request message is received (timing accumulated in a buffer) when no queuing delay occurs.
The arrow pointing upward indicates the timing at which the Delay_Request message packet is actually received (the timing at which the packet is actually stored in the buffer).
When a packet of a Delay_Request message is received, the amount of accumulation in the packet buffer is increased by the number of received bits. When the vertically extending broken line and the upward arrow coincide with each other on the time axis, that is, when a Delay_Request message packet is received without causing a queuing delay, the read timing (accumulation amount is The time width between the timing (decrease timing) and the newly received timing (timing when the accumulation amount increases) is the minimum Tmin. On the other hand, when a queuing delay occurs, the above time width becomes a value larger than Tmin. At this time, the difference between the minimum value Tmin and the time width is equivalent to the delay amount of the generated queuing delay.

Note that any of the configurations of the delay measurement unit 403 described above is merely an example of a configuration for measuring a delay, and other configurations may be employed as long as the configuration can measure the delay amount MS_Q of the queuing delay. .

As described above, the embodiment of the present invention has been described in detail with reference to the drawings. However, the specific configuration is not limited to this embodiment, and includes a design and the like within the scope not departing from the gist of the present invention.

According to the present invention, it is possible to realize highly accurate time synchronization with reduced costs in mobile backhaul, factory automation, home AV LAN, and the like.

300 ... Master node, 301 ... Clock generation unit, 302 ... Master clock unit, 303 ... Packet generation unit, 304 ... Packet transmission unit (transmission unit), 305 ... Packet reception unit, 400 ... Slave node, 401 ... Packet reception unit, 402: PLL unit, 403: Delay measurement unit (delay measurement unit), 404 ... Slave clock unit, 405 ... Time synchronization control unit, 406 ... Packet transmission unit (transmission unit), 407 ... Phase comparator, 408 ... LPF, 409 ... PI controller, 410 ... VCO, 411 ... counter, 501 ... packet counter, 502 ... counter maximum value monitoring unit, 503 ... counter value monitoring unit upon arrival, 504 ... delay calculation unit, 701 ... packet analysis unit, 702 ... packet Generation unit, 703 ... RTT calculation unit, 70 ... propagation delay monitor (propagation delay measurement unit), 705 ... offset calculation section, 706 ... time adjustment unit

Claims (8)

A time synchronization system that includes a master node and a slave node that communicate with each other, and synchronizes the time at the slave node with the time at the master node,
The master node is
A transmission unit for transmitting a control message to the slave node;
A receiver for receiving a control message from the slave node;
With
The slave node is
A transmitter for transmitting a control message to the master node;
A receiving unit for receiving the control message from the master node;
A delay measuring unit that measures a delay amount representing a queuing delay received in a communication path by a control message transmitted from the master node to the slave node;
A propagation delay measuring unit for measuring a propagation delay amount between the master node and the slave node;
The difference between the time at the slave node and the time at the master node is calculated using the delay amount measured by the delay measurement unit and the propagation delay amount measured by the propagation delay measurement unit, and the slave node A time synchronization controller that synchronizes the time at the master node with the time at the master node;
A time synchronization system comprising: The transmitting unit and the receiving unit of the master node and the transmitting unit and the receiving unit of the slave node each repeatedly transmit and receive the control message,
The propagation delay measurement unit detects a round-trip transmission delay amount in transmission / reception of the control message repeated within a predetermined period, and calculates a difference between the round-trip transmission delay amount and the delay amount measured by the delay measurement unit. The time synchronization system according to claim 1, wherein the propagation delay amount is measured based on a minimum value. The delay measurement unit measures the delay amount based on the accumulated amount of packets when the control message is received and the maximum accumulated amount of packets in the immediately preceding monitor period for each predetermined monitoring period. The time synchronization system according to claim 1. The delay measuring unit includes a counter that counts a current packet accumulation amount of the control message, and the counter increases a fixed value every time the control message is received, and the counter of the accumulated control message 4. The time synchronization system according to claim 3, wherein the counter value decreases at a constant pace by reading processing. The delay measuring unit includes a buffer for accumulating the received control message, and based on a timing at which the control message is accumulated in the buffer and a timing at which the control message is periodically read from the buffer. The time synchronization system according to claim 1, wherein the delay amount is measured. A slave node that synchronizes time with a master node that includes a transmitter that transmits a control message to a slave node and a receiver that receives a control message from the slave node,
A transmitter for transmitting a control message to the master node;
A receiving unit for receiving the control message from the master node;
A delay measuring unit that measures a delay amount representing a queuing delay received in a communication path by a control message transmitted from the master node to the slave node;
A propagation delay measuring unit for measuring a propagation delay amount between the master node and the slave node;
The difference between the time at the slave node and the time at the master node is calculated using the delay amount measured by the delay measurement unit and the propagation delay amount measured by the propagation delay measurement unit, and the slave node A time synchronization controller that synchronizes the time at the master node with the time at the master node;
A slave node comprising: A time synchronization method comprising a master node and a slave node that communicate with each other, and performed by a time synchronization system that synchronizes the time at the slave node with the time at the master node,
A transmission step in which the master node transmits a control message to the slave node;
A receiving step in which the master node receives a control message from the slave node;
A transmission step in which the slave node transmits a control message to the master node;
A receiving step in which the slave node receives the control message from the master node;
A delay measuring step in which the slave node measures a delay amount representing a queuing delay received in a communication path by a control message transmitted from the master node to the slave node;
A propagation delay measurement step in which the slave node measures a propagation delay amount between the master node and the slave node;
The slave node calculates a difference between the time at the slave node and the time at the master node using the delay amount measured at the delay measurement step and the propagation delay amount measured at the propagation delay measurement step. A time synchronization control step for synchronizing the time at the slave node with the time at the master node;
A time synchronization method comprising: As a time synchronization system that includes a master node and a slave node that communicate with each other and synchronizes the time at the slave node with the time at the master node, a first computer that corresponds to the master node and a second computer that corresponds to the slave node A time synchronization program for operating,
For the first computer,
A transmission step of transmitting a control message to the slave node;
Receiving a control message from the slave node;
And execute
For the second computer,
A transmission step of transmitting a control message to the master node;
Receiving the control message from the master node;
A delay measuring step of measuring a delay amount representing a queuing delay received in a communication path by a control message transmitted from the master node to the slave node;
A propagation delay measuring step for measuring a propagation delay amount between the master node and the slave node;
Using the delay amount measured by the delay measurement step and the propagation delay amount measured by the propagation delay measurement step, the difference between the time at the slave node and the time at the master node is calculated, and the slave node A time synchronization control step of synchronizing the time at the master node with the time at the master node;
The program for time synchronization characterized by performing this.

Images