JP2020107943A - Time division schedule adjusting method - Google Patents

Abstract

To prevent decrease in a transmission path utilization rate by delivering a packet to a transmission destination within a time division period without extending the time division period in the case where a transmission distance between communication apparatuses is long.SOLUTION: Time synchronization is performed among HUB-A to HUB-D on a network, and the network is time-divided into each of the HUB-A to HUB-D. This time division schedule is offset according to the distance between the HUB-A to HUB-D. The values of offset F1 to F3 are set for each port of the HUB-A to HUB-D.SELECTED DRAWING: Figure 4

Description

The present invention relates to a technique of adjusting a schedule of a network time-divided by TSN (Time Sensitive Networking).

In order to protect important communication in the network, priority control and bandwidth control are often used. Even if the functions such as the priority control and the band control are used, in a situation where a plurality of transmissions are momentarily overlapped, the transmissions may be discarded and may not reach.

As a result, it is difficult for the control such as the priority control and the bandwidth control to ensure the reliable arrival of the transmission in a complicated network. Therefore, a standard called TSN of Non-Patent Document 1 is being established in "IEEE 802.1".

According to TSN, each communication device performs time synchronization by PTP (Precision Time Protocol) of Non-Patent Document 2, and time-divisionally adjusts the network so that transmission of important data surely arrives.

Further, in the TSN, since important packets and general packets are time-divided, the transmission timing is controlled at the transmission port in transmission queue units. Here, there is a time-divided schedule, and the schedule is determined in consideration of not only the specifications required by the system but also the transmission delay and the detour function due to the network configuration.

"Outline and implementation of "TSN" that fuses IT and OT networks", [online], search November 22, 2018, Internet <URL: http://ednjapan.com/edn/articles/1803/23/ news014.html> "Endrum IEEE1588 PTP Grandmaster Clock", [online], October 15, 2018 search, Internet <URL: https://www.shoshin.co.jp/c/endrum/1588ptp.html>

The TSN described above performs time synchronization by PTP to synchronize the clocks of the communication devices on the network, time-divisions the network by the clock, and transmits important data and general data separately. At this time, since the important data can be occupied in a predetermined time zone, it can be delivered to the other party without being affected by the general data.

For example, TNS is used for automatic driving of automobiles. In this automatic driving, an accident may occur if a large amount of data from various sensors cannot be delivered in time. In addition, dedicated transmission lines and interfaces are used for high-speed control of industrial equipment, etc., and TSN may be applied to these, but such systems are limited to vehicles, factory premises, and premises, so communication distance Is short and the transmission delay is small. However, if the TSN is mounted on a network having a long transmission line, the transmission delay is large and the utilization efficiency may be deteriorated.

For example, the transmission speed of an optical fiber or a copper-wired Ethernet (registered trademark) cable is “200,000,000 m/sec”, so that a transmission delay of [100 μsec] occurs when the transmission line is 20 km. Further, if the transmission rate is "1 Gps", the transmission time of "1518 bytes" will be "12.3 µsec". Therefore, the time division period of TSN becomes longer than the communication time due to [transmission delay+transmission time].

Referring to FIG. 1, the network 1 includes four communication devices, that is, HUB-A to D (switching hubs) as relay devices, and the HUB-A to D are TSN compatible devices. Since the distances between the HUBs A to D are greatly different from each other, there is a large difference in [transmission delay]. This point will be described by taking the transmission direction R (communication direction, transfer direction) in FIG. 1 as an example.

FIG. 2 shows a state in which all of HUB-A to D are time-shared at the same time by the time synchronization by PTP. Here, when the packet 2 is transmitted from the HUB-A to the HUB-D, the influence of “transmission delay” is great, and it takes time for the packet 2 to reach the HUB-D. Therefore, the time-divided period (time) T is set. Making it big.

At this time, if it is attempted to deliver “1518 byte 5 packets” to HUB-D 100 km away from HUB-A, it takes [transmission delay=500 μsec]·[transmission time=61.5 μsec]. That is, it takes “561.5 μsec” to transmit “61.5 μsec”, and the utilization factor of the transmission path is about 11%. Further, since the transfer time of each HUB-A to D, the guard band 5, and the error of time synchronization are taken into consideration, the actual utilization factor of the transmission path further decreases.

As a result, as shown in FIG. 2, when the transmission distance D is long, the time-division period 4 becomes large, so that the update cycle in the system becomes slow. Therefore, TSN may not be applicable to a system that needs to be updated in a short time. Further, even if the transmission distance D is short, the influence of [transmission delay] is relatively large in a system having a fast update cycle, and thus TSN may not be applicable in the same manner.

Here, in FIG. 2, each HUB-A to D simultaneously transmits the packet 2 at the beginning of the time division period 4. That is, although it is considered that the transfer via the HUB-B and C is performed in the middle of communication, when a packet is transmitted from the terminal devices belonging to the HUB-A to D in the middle of the time division period 4, it is shown in FIG. As described above, there is a possibility that the other party arrives outside the time-sharing period 4.

Outside this period 4, general communication other than important data is being carried out, so that it is not possible to guarantee that the HUB-D will be reached. Then, (1) not only the time division period 4 is lengthened, but communication other than the transfer from other HUB-B, C is limited to only the period determined by the design of the period 4, and (2) the destination. It is necessary to calculate the arrival time and send it only when it is within the range of period 4, but it is difficult to realize it.

The present invention has been made to solve such a conventional problem, and when the transmission distance between communication devices is long, the packet is delivered to the destination within the time division period without increasing the time division period. The problem to be solved is to suppress the decrease in the utilization factor of the transmission line.

(1) The present invention is a method for synchronizing time between communication devices on a network, time-sharing the network for each communication device, and adjusting the time-sharing schedule.
The time-division schedule is offset according to the distance between the communication devices. The offset is, for example, a portion of F1 to F3 in FIG.

(2) One aspect of the present invention is characterized in that the offset value can be set for each port of the communication device.

(3) Another aspect of the present invention is a relay device in which the communication device relays transmission and reception of packets between terminals,
The survey packet is transmitted from the relay device designated as the end of the network,
It is characterized in that the value of the offset is adjusted according to the reception timing of the investigation packet in another relay device.

(4) Yet another aspect of the present invention is characterized in that the investigation packet is transmitted by unicast, broadcast, or multicast.

(5) In still another aspect of the present invention, the relay device calculates a time difference between the time division schedule and the reception timing of the investigation packet,
The time difference is recorded as a record in the offset table of the transfer destination port of the investigation packet,
The smallest value of the time difference among the records of the offset table is set as the value of the offset.

(6) Yet another aspect of the present invention is characterized in that a warning is given if there is a record in the offset table outside the schedule period.

According to the present invention, when the transmission distance is long, the packet is delivered to the destination within the time-division period without increasing the time-division period. Therefore, it is possible to suppress the decrease in the utilization factor of the transmission path.

Network configuration diagram. FIG. 2 is a state diagram in which the network of FIG. 1 is time-divided by TSN. FIG. 3 is a state diagram in which packets are transmitted during the time division period of FIG. 2. FIG. 6 is a time division timing chart showing the time division schedule adjustment method according to the embodiment of the present invention. The state diagram in which the investigation packets of the embodiment are sequentially transferred. Explanatory drawing of the value of the offset of the same transmission port.

Hereinafter, a method of adjusting the time division schedule according to the embodiment of the present invention will be described. This adjustment method adjusts a schedule in which the network is time-divided between communication devices by TSN on the assumption that each communication device on the network is time-synchronized by PTP.

In the conventional system using TSN, the following problems (1) to (3) may occur.
(1) When the transmission distance between communication devices becomes long, the transmission time becomes long, and the time division period (time) must be lengthened.
(2) If the time-division period is long, the utilization factor of the transmission line may be significantly reduced.
(3) If the time division period (time) is long, the number of time divisions is reduced, and it becomes difficult to apply it to a communication system with a short update cycle.

Therefore, when the transmission distance between the HUBs is long, the adjustment method has been proposed in which the packet reaches the destination within the time division period without increasing the time division period. An example of applying the adjustment method to the network 1 of FIG. 1 will be described with reference to FIG.

FIG. 4 shows a time division schedule when the packet 2 is transmitted (transmitted) in the transmission direction R from the HUB-A as in FIG. Here, the time division timing is shifted by the offset F in consideration of the transmission direction R and the distance D of the packet 2.

At this time, if the transfer of each HUB-A to D is sequentially processed in a pipeline, the influence of [transmission delay] can be minimized. Further, the time-division period 4 is minimized, the utilization factor of the transmission path is improved, and the update cycle is shortened.

Furthermore, a method for preventing new transmission in the middle of the time division period 4 is required, but since the time division period 4 can be shortened by the above-mentioned offset adjustment method, a simple guard band 5 can be used. The method can be applied.

1 to 4 describe the transmission in the transmission direction R, the same applies to the transmission in the transmission direction L and the time division is performed at a timing different from that in the transmission direction R. At this point, the time-division timing schedule is performed for each transmission direction, that is, for each port (unit of port).

<<Example>>
An embodiment of the adjusting method will be described with reference to FIGS. 5 and 6. In this embodiment, the survey packet 12 is used to automatically adjust the value of the offset F of the time division schedule.

That is, since the network 1 of FIG. 1 has a simple configuration, only the communication directions R and L are shown. However, the actual network configuration is complicated, and there are also detouring functions such as STP (Spanning Tree Protocol)/RSTP (Rapid Spanning Tree Protocol) and link structures designed to prevent loops.

In order to understand the structure of such a complicated network, it is possible to collect information of adjacent HUBs and automatically determine the time division period (time) and offset value of TSN from the measured values by PTP. I think that the.

However, although protocols such as STP/RSTP can interpret the tree structure at present, they cannot recognize the configuration of the entire network, and it is easy to solve the determination of the time division period and the offset value. Not.

At this time, according to the network redundancy function, another communication path becomes effective in the event of a line failure, but since the network connection configuration and distance change, the offset value must be changed. In addition, since it detours to another communication path, another port is used and the offset value can be used for the corresponding port, but the offset value cannot be changed to the network beyond that.

Therefore, in view of such a situation, the adjustment method is proposed in which the offset value of the time division schedule is automatically adjusted according to the reception timing of the investigation packet.

(1) Survey packet First, the survey packet will be described with reference to FIG. FIG. 5 shows a time division schedule of the network 1 of FIG. 1, as in FIGS. 2 and 3. Here, for convenience of comparison and comparison, description will be made based on the network 1 in FIG. 1, but the same applies to other networks.

FIG. 5 shows a state in which the investigation packet 12 is transmitted in the communication direction R from the relay device designated as the end of the network 1, that is, the HUB-A. Further, although omitted in FIG. 5, since the HUB-D is also a relay device at the end of the network 1, if specified, the investigation packet 12 is similarly transmitted.

At this time, the investigation packet 12 is transmitted with a fixed-length data amount from the beginning of the time division period of TSN, and information of the time of the time division period by TSN/number indicating the time is added.

This investigation packet 12 matches the destination, unicast/broadcast/multicast type, etc. according to the important packet transmitted by the terminal devices T1, T2 (see FIGS. 1 and 6) directly connected to the HUB-A.

For the inspection packet 12, a method of using the time stamp mechanism of PTP and a method of transmitting in the time-divided period for the inspection packet 12 are used.

In the case of the method using the time stamp mechanism, it may be transmitted during the period of a general packet. It is received and transferred from HUB-A to HUB-D while adding the time elapsed on the transmission path. The added value at this time is set as the value of the offset F of each HUB and recorded in the offset table 13 (see FIG. 6) provided for each transfer destination port.

In the case of such a method that does not use the time stamp mechanism, it is preferable that a time-divided period for the inspection packet 12 is provided because it is affected by an important or general packet. In this method, when the survey packet 12 is received, it is recorded in the offset table 13 by referring to the internal clock for time synchronization.

The survey packet 12 is received and transferred by the TSN compatible device. Here, the investigation packet 12 transmitted from HUB-A is received and transferred to HUB-B to D. At this time, the time difference between the time division schedule (offset value=0) and the reception timing of the investigation packet 12 is calculated.

The calculation result is recorded in the offset table 13 (see FIG. 6) provided for each transfer destination port. At this time, since the time difference is recorded in the offset table 13 every time the survey packet 12 is transmitted, a plurality of records are registered in the offset table. The offset value is adjusted to the timing closest to the time of the time division schedule (offset value=0) in this record group.

(2) Processing Content Next, processing content of the embodiment (S01 to S06 not shown) will be described.

S01: First, the system designer determines a time division schedule and a cycle in consideration of the amount of packets flowing through the network 1 and the number of stages of HUB-A to D.

S02: Next, the system designer designates a plurality of HUBs to which terminal devices for transmitting and receiving important packets are directly connected. Here, the end HUB-A, D is designated for each of the communication directions R, L.

S03: The HUB-A, D designated in S02 transmits the survey packet 12 at the beginning of the period if the time-divided period for the survey packet 12 is provided. On the other hand, in the case of using the time stamp mechanism, the survey packet 12 is transmitted during the period of the general packet.

At this time, as described above, the data amount of the investigation packet 12 has a fixed length, and includes the time of the time-divided period/the number indicating the time. Further, the investigation packet 12 is transmitted by any one of unicast, broadcast and multicast as with the important packet.

Note that the destination of the unicast transmission of the investigation packet 12 is the HUB to which the terminal device of the destination of the important packet is directly connected. For example, when the important packet is transmitted from the terminal devices T1 and T2 to the terminal devices T3 and T4 (see FIGS. 1 and 6) by unicast, the inquiry packet 12 is transmitted by unicast from HUB-A to HUB-D. It

S04: The HUB-A to D that have received the check packet 12 of S03 use the time stamp mechanism as a new offset value obtained by obtaining the transmission time from the time stamp and adding it to the offset value in the packet. , Is recorded and registered in the offset table 13 for each transfer destination port.

On the other hand, in the case of the method that does not use the time stamp mechanism, the time difference between the time division schedule (offset value=0) and the reception timing of the investigation packet 12 is calculated for each of the communication directions R and L. The time difference calculated here is recorded and registered in the offset table 13 for each transfer destination port of the investigation packet 12.

Table 1 shows the contents of one record in the offset table 13. Here each record is
-The number of the receiving port of the investigation packet 12-The MAC address of the sender of the investigation packet 12-The number of seconds that have passed since the investigation packet 12 was received (counted down and erased with a count of "0")
-Time division time/number of the time-Time division timing (offset value = "0") and the time difference between the reception of the survey packet 12 are provided.

At this time, since there is only one transfer destination of the investigation packet 12 by unicast, it may be registered in the offset table 13 of the corresponding port. On the other hand, in the case of broadcast or multicast, it is registered in the offset table 13 of all ports other than the receiving port which is the same VLAN.

For example, as indicated by an arrow U in FIG. 6, the time difference based on the investigation packet 12a for unicast transmission is registered only in the offset table 13b. On the other hand, as indicated by the arrow B, the time difference based on the survey packet 12b for broadcast transmission is registered in the offset tables 13a to 13d.

S05: A record of the time difference based on the survey packet 12 received in the past is registered in each offset table 13. That is, in each offset table 13, a plurality of records of the time difference according to the number of received survey packets 12 are registered.

Therefore, the smallest time difference in each offset table 13 is adopted as the offset values F1 to F3, and the time division schedule is offset based on the offset values F1 to F3. As a result, the time division schedule is offset for each of the communication directions R and L, and the following effects (A) to (F) are obtained.

(A) That is, in the conventional TSN, when the HUBs A to D are time-shared at the same time in time synchronization by PTP, the time-division period is increased due to the influence of [transmission delay] as shown in FIG. It was necessary to set the communication to end within the specified period.

On the other hand, according to the adjustment method, the survey packet 12 is transmitted from each of the HUB-A and D designated as the terminals of the communication directions R and L. Therefore, the time division schedule is offset for each of the communication directions R and L, and the length of the time division period can be suppressed to the necessary minimum.

(B) Further, if there is a change in the configuration of the network 1, the survey packet 12 is transmitted to follow the change, and the offset values F1 to F3 are adjusted so that the survey packet 12 is included in the time division schedule. To be done.

In this case, since the network configuration is complicated, it is possible to set an appropriate offset by similarly distributing the survey packets 13 so that all the important packets are accommodated in the time-divided period.

(C) In the conventional TSN, the longer the transmission distance between the HUBs, the longer the transmission time. Therefore, it is necessary to lengthen the time division period. On the other hand, according to the adjustment method, the length of the time-divided period can be suppressed to the necessary minimum as described above.

(D) TSN tends to significantly reduce the utilization factor of the transmission path when the time division period is long. In this respect, according to the adjusting method, the length of the time-divided period can be suppressed to the necessary minimum as described above, so that it is possible to prevent the utilization rate of the transmission line from decreasing.

(E) If the time-division period is long, the number of time-divisions is reduced, making it difficult to apply the system to a system having a short update cycle. With respect to this point, according to the adjustment method, the length of the time-divided period can be suppressed to a necessary minimum as described above, and therefore, it can be applied to a system having a short update cycle.

(F) The TSN needs to change the offset value when the network configuration changes or when the bypass function such as STP/RSTP is used. In this respect, the adjustment method can automatically adjust the offset value of the time division schedule by retransmitting the survey packet 12 without requiring recognition of the network configuration.

S06: In S05, it is confirmed whether or not the time difference not adopted as the offset value is within the period of the time division schedule. That is, after the minimum value in the offset table 13 is adopted as the offset value in S05, it is confirmed whether or not the remaining time difference registered in the offset table 13 is within the period in the time division schedule.

At this time, if the offset table 13 has the time difference outside the period of the time division schedule, it is assumed that the system setting is incorrect or the network 1 is in an unexpected state.

Therefore, if such time difference exists in the offset table 13, it is recorded as a warning. Then, although it cannot be confirmed whether the important packet is received within the time-division period, the presence or absence of the warning can be used as an index for normal operation of the TSN.

It should be noted that the TSN cannot operate properly when the inspection packet 12 does not arrive or when the data in the offset table 13 remains. Therefore, each HUB-B, C receives the survey packet 12 and overwrites it on the offset table 13, but when reception of the survey packet 12 is interrupted, it is deleted by a timeout.

1... Network 2, 3... Packet 4... Time-divided period 5... Guard band 12... Survey packet 13... Offset table

Claims (6)

A method for synchronizing time between communication devices on a network, time-sharing the network for each communication device, and adjusting a schedule of the time-sharing,
A time-division schedule adjustment method, wherein the time-division schedule is offset according to the distance between the communication devices. 2. The time division schedule adjustment method according to claim 1, wherein the offset value can be set for each port of the communication device. The communication device is a relay device that relays transmission and reception of packets between terminals,
The survey packet is transmitted from the relay device designated as the end of the network,
3. The time division schedule adjustment method according to claim 1, wherein the offset value is adjusted according to the reception timing of the investigation packet in another relay device. The time division schedule adjustment method according to claim 3, wherein the investigation packet is transmitted by unicast, broadcast, or multicast. The relay device calculates a time difference between the time division schedule and the reception timing of the investigation packet,
The time difference is recorded as a record in the offset table of the transfer destination port of the investigation packet,
5. The time division schedule adjustment method according to claim 3, wherein the minimum value of the time difference among the records in the offset table is set as the offset value. 6. The time-division schedule adjustment method according to claim 5, wherein a warning is issued if there is a record in the offset table outside the schedule period.

Images