JP7380359B2 - Redundant route management method, redundant route management method - Google Patents

Description

The present invention mainly relates to a network with redundant master and slave routes using PTP-compliant communication equipment.

PTP (Precision Time Protocol) is a relatively new protocol created to obtain highly accurate time synchronization by restricting the usage environment to a LAN, and the specifications of PTP are defined as IEEE1588.

In "PTP", a "Grandmaster (GM)" performs sophisticated time distribution, and a slave receives the time. In addition, "PTP" uses a hardware stamp function implemented in the "MAC" and "PHY" network interface chips to achieve timestamp accuracy of microsecond RMS or less.

Currently, there are two versions of the "PTP" protocol, and "Version 1" has a "Boundary Clock (BC)" that separates segments for large-scale deployment.

"Version 2" includes a switching hub "Transparent Clock (TC)" with a delay management function, enabling more flexible and highly accurate deployment.

A communication device incorporating this PTP function, that is, an L2 (layer 2) switching hub (hereinafter referred to as a PTP hub) is capable of highly accurate time synchronization with the GM. In addition, by connecting PTP hubs and exchanging PTP packets between them, a master is automatically selected from among the PTP hubs (master selection process), and the time of the PTP hub of the selected master is PTP hubs can act as slaves and perform time synchronization.

The master PTP hub transmits PTP packets to other PTP hubs at regular intervals to maintain time synchronization. In addition, all PTP hubs have an original function that allows them to notify non-PTP peripherals of status information (PTP information) inside the PTP hub at regular intervals as PTP information packets.

JP2018-174445

A virtual LAN (VLAN) sets up a virtual group of terminals in a corporate network (LAN) independently of their physical connection form. Here, the functions of a communication device (hub: HUB) called a LAN switch are used to group the terminals according to the protocols used, such as the MAC addresses and IP addresses of the terminals. This has the advantage that the network configuration can be changed regardless of the physical location of the terminal, and there is no need to change settings even if the terminal is moved.

By using a hub (hereinafter referred to as a PTP hub) having a redundancy function of VLAN duplication using PTP VLAN, redundancy by VLAN duplication becomes possible. At this time, the PTP hubs have a route switching function that selects and connects a route with good conditions from among the duplicated routes. Patent Document 1 is known as a technology that applies such PTP time synchronization to a virtual LAN.

However, hubs that cannot support VLAN duplication between PTP hubs (hubs that have a PTP function but do not have a VLAN duplex function/hubs that have a VLAN duplex function but the assumed redundancy through VLAN duplexing does not function sufficiently) If this exists, there is a risk that the time synchronization process will be adversely affected.

In this case, if one route suddenly deteriorates, switching to another route is performed due to redundancy. However, if the time synchronization accuracy between PTP hubs deteriorates or if PTP packets stop being transferred midway through, it may take a considerable amount of time to switch to another route, or PTP processing for one or both routes may be delayed. may become impossible to execute.

The present invention has been made to solve such conventional problems, and its object is to detect route abnormalities in redundant VLANs at an early stage and take appropriate measures.

(1) One aspect of the present invention is a network management method that provides redundant master-slave paths between time-synchronized masters and slaves,
"Meanpath" is calculated as an average value of the transmission time of the outgoing path and the incoming path between the master and the slave, and if the calculated "Meanpath" continues to be in a negative state, it is determined that the time stamp is abnormal;
The present invention is characterized in that the master/slave of the route is switched depending on the determination.

(2) Another aspect of the present invention is a method for managing a network in which time-synchronized master and slaves are made redundant using master-slave routes, comprising:
a step of calculating "Meanpath" as an average value of the transmission time of the outbound path and the return path between the master and the slave;
determining that the timestamp is abnormal if the calculated "Meanpath" continues to be in a negative state;
switching the master and slave of the route according to the determination;
It is characterized by having the following.

According to the present invention, it is possible to detect route abnormalities in redundant VLANs at an early stage and take appropriate measures.

A network configuration diagram showing an example of VLAN duplex connection (normal) of a PTP hub, A state diagram of an abnormality occurring in a part of the route in Figure 1, FIG. 7 is another network configuration diagram showing connections for VLAN duplication. FIG. 3 is a state diagram in which a PTP hub has a VLAN duplexable area and an unduplexable area. 5 is a flowchart showing the overall processing of the redundant route management method according to the embodiment of the present invention. 6 is a flowchart showing details of S09 in FIG. 5. 6 is a flowchart showing details of S10 in FIG. 5. 8 is a flowchart showing details of S37 in FIG. 7. 9 is a flowchart showing details of S47 in FIG. 8. 7 is a flowchart showing details of S28 in FIG. 6 and S39 in FIG. 7.

The details of the redundant route management system (method) according to the embodiment of the present invention will be described below. Here, we assume a network using a plurality of PTP hubs, that is, a network with redundant routes through VLAN duplication, and appropriately deal with abnormalities by managing each route.

Explaining based on FIG. 1, the route between PTP hubs 1 and 2 is duplicated by VLANs 1 and 5, with the VLAN 1 side being called the A route and the VLAN 5 side being called the B route. Redundancy is provided between the PTP hubs 1 and 2 through routes A and B, and PTP time synchronization is performed using either route.

At this time, the PTP hub 1 is in the master state, the PTP hub 2 is in the slave state, and the A route is the main VLAN (main route), while the B route is operating as the sub VLAN (subordinate route). "V2-M-Aroot" in FIG. 1 indicates the A route as the main root, and "V5-S-Broot" indicates the B route as the sub-VLAN.

The main VLAN mainly performs evaluation of route A (hereinafter referred to as evaluation A) by PTP processing and time synchronization. Further, the sub-VLAN mainly executes only evaluation of route B (hereinafter referred to as evaluation B) by PTP processing. Here, if the disconnection/evaluation A of the A route is worse than the evaluation B and the switching condition is satisfied, the B route is switched to the main VLAN, and the A route is switched to the sub VLAN.

Hubs A to D interposed between PTP hubs 1 and 2 have a PTP function, but cannot support VLAN duplication. For example, it cannot be transmitted when connecting with "EtoE" / It cannot support multiple domains / Time stamp (hereinafter referred to as "TS") processing methods are different / Only step 2 is supported / Hubs with malfunctions or failures, etc. do.

≪Processing details of redundant route management method (method)≫
(1) Redundancy processing when VLAN is used The redundancy processing when VLAN is used (S01 to S11), which is the main part of the redundant route management method, will be explained based on FIG. S01 to S8 and S11 during this process are executed by both the master PTP1 and slave PTP2,
Here, it is assumed that initial information A to C is set in advance by the user's command input and written to the initial setting file FA1.
A: Information for determining whether it is necessary to change from multicast to broadcast for each port (hereinafter referred to as port BC conversion necessity information D1).
B: Number of abnormality determinations until a final determination is made that an abnormality occurs when an abnormality related to time synchronization occurs (hereinafter referred to as the TS abnormality determination number D2).
C: The number of normality judgments until the final judgment is made as normal at the time of the above judgment (hereinafter referred to as the TS normality judgment number D3).
S01, S02: Processing is started by starting the PTP hubs 1 and 2. At the start of processing, it is checked whether initial settings are required. As a result of this confirmation, if it is not necessary, the process is terminated, while if necessary, the information of port BC conversion necessity information D1, TS abnormality judgment count D2, and TS normality judgment count D3 is read from the initial setting file FA1. It is set (S02).

Thereafter, conventional PTP hub processing is executed, and PTP hub 1 is selected as the master by the master selection processing described above, while PTP hub 2 is synchronized with PTP hub 1 as a slave. The processes of S03 to S08 and S11 are executed as necessary until the state changes to either the master state or the slave state. After the state transition to master/slave, processes S04 to S11 are executed.

S03: Initialize the abnormality counter (counts the number of abnormality determinations) D4, the normality counter (counts the normality determinations), the switching lock flag F1, and other necessary information.

S04 to S06: In order to check whether it is necessary to receive not only the original multicast packet but also the broadcast packet, the BC conversion necessity information of the port is checked (S04). As a result of the confirmation, if necessary, the process proceeds to S05 to execute processing for receiving broadcast packets (S05), while if not necessary, proceeding to S06 to execute conventional processing for receiving multicast packets (S06).

S07 to S11: As a result of the master selection process, it is confirmed whether or not the master state is reached (S07). As a result of the confirmation, if it is in the master state, the process advances to S09 and executes PTP packet transmission processing (master).

On the other hand, if it is not in the master state, the process advances to S08 to check whether it is in the slave state. As a result of the confirmation, if it is in the slave state, the PTP packet transmission process (slave) of S10 is executed.

In the network configuration of FIG. 1, since the PTP hub 1 is in the master state, it executes the PTP packet transmission process in S09. On the other hand, since the PTP hub 2 is in the slave state, it executes the PTP packet transmission process in S10. Note that, as a result of the confirmation in S08, if the device is not in the slave state, the conventional PTP process, that is, the master selection process, etc. is executed (S11).

(2) PTP packet transmission processing (master)
Details of the PTP packet transmission process in S09 will be explained based on FIG. 6. This process is executed by the PTP hub 1 that is the master.

S21: When the process is started, it is checked whether the received packet P is a PTP process reset preparation request. As a result of the confirmation, if it is a PTP processing reset preparation request, the process proceeds to S22, whereas if it is not a PTP processing reset preparation request, the process proceeds to S23.

S22: A preparation flag for resetting the PTP processing (PTP processing reset preparation flag F2) is enabled. Here, the flag F2 is set to "valid=1" and the process proceeds to S29.

S23: It is confirmed whether the received packet P is a PTP processing reset start request. As a result of the confirmation, if it is a PTP processing reset start request, the process proceeds to S24, whereas if it is not a PTP process reset start request, the process proceeds to S26.

S24: Check the PTP processing reset preparation flag F2, and if the flag F2 is valid, proceed to S25, while if invalid, proceed to S29.

S25: Execute PTP reset processing similar to the conventional one.

S26: For processing other than packet transmission, conventional PTP processing is executed.

S27 to S29: It is checked whether the PTP packet can be transmitted (S27). As a result of the confirmation, if transmission is possible, the process advances to S28.

In S28, the original multicast transmission is converted to broadcast transmission, and the PTP packet is transmitted from the PTP port. On the other hand, if transmission is not possible, the process advances to S29, where conventional PTP timer processing is executed, and after execution, the process is terminated.

(3) PTP packet transmission processing (slave)
The details of the PTP packet transmission process in S10 will be explained based on FIG. 7. This process is executed by the PTP hub 2 serving as a slave.

S31-S36: Processes similar to S21-S26 are performed.

S37: TS abnormality processing is performed.

S38-S40: Processes similar to S27-S29 are performed.

(4) TS abnormality processing The TS abnormality processing in S37 will be explained based on FIG. 8. Here, "Meanpath" is used to determine the occurrence of a TS abnormality.

That is, the average value of the transmission time between the master and the slave on the outward and return paths is called "Meanpath." This "Meanpath" value usually does not become negative, but it may occur temporarily depending on the system. However, if it continues for a long time, it is assumed that some factor causing a TS abnormality between the PTP hubs 1 and 2 (it may be an abnormality in the TS processing of the PTP hub) has occurred.

Therefore, if left as is, the time synchronization will gradually deviate, and there is a risk that a synchronization abnormality will occur after several hours, and a certain deviance will continue. In such a case, it is necessary to quickly deal with the cause of the abnormality by switching routes without using the route where the TS abnormality occurred.

S41: When the process is started, it is confirmed whether or not a TS abnormality has occurred. Specifically, when "Meanpath" is negative, it is confirmed that a TS abnormality has occurred. As a result of this confirmation, if an abnormality has occurred, the process proceeds to S42, whereas if no abnormality has occurred, the process proceeds to S48.

S42, S43: First, "1" is added to the abnormality counter D4 (S42). It is checked whether the value of the abnormality counter D4 after this addition exceeds the TS abnormality determination number D2 (S43). As a result of the confirmation, if the amount has exceeded, the process proceeds to S44, whereas if the amount has not exceeded, the process ends.

S44 to S47: The switching lock flag F1 of the route to be determined, that is, the route in which the TS abnormality has occurred among routes A and B, is set to "valid=1" (S44). As a result, the route to be determined is finally determined to have a TS abnormality, and the content of flag F1 "valid = 1" is included in the PTP information packet (sent at regular intervals between PTP hubs 1 and 2). The occurrence of the TS abnormality is notified to external equipment by means of (S45).

Further, the master of the route to be determined, that is, the PTP hub higher than the PTP hub 1, is notified of reset preparation (S46). After this notification, master/slave VLAN switching determination processing is executed (S47).

S48: Check whether the switching lock flag F1 is valid or invalid. As a result of the confirmation, if it is valid, the process proceeds to S49, whereas if it is invalid, the process ends.

S49, S50: Add "1" to the normal counter D5 (S49). It is checked whether the value of the normality counter D5 after this addition exceeds the TS normality determination number D3 (S50). As a result of the confirmation, if the amount has exceeded, the process proceeds to S51, whereas if it has not exceeded, the process ends.

S51, S52: Clear the abnormal counter D4 and the normal counter D5 (S51), and set the switching lock flag to "invalid=0" (S52). As a result, the route to be determined is finally determined to be normal, and the process ends thereafter.

(5) Master/Slave VLAN Switching Determination Process The details of the VLAN switch determination process in S47 will be explained based on FIG. 9.

S61: When the process starts, the A route evaluation value AD and the B route evaluation value BD are obtained based on the information of the switching lock flag F1, the master collection data D6, and the slave collection data D7. At this time, if the information of the switching lock flag F1 is "valid=1", the evaluation values of both the evaluation values AD and BD are set to "0". This is taken as the minimum evaluation value.

S62: Check whether the A route evaluation value AD is smaller than the route evaluation allowable evaluation value RD. As a result of the confirmation, if it is smaller, the process proceeds to S63, whereas if it is not smaller, the process is ended. At this time, if the information of the switching lock flag F1 is "valid = 1", the evaluation value is the minimum as described above, so it is smaller than the allowable evaluation value RD.

S63, S64: Calculate the difference between "(B route evaluation value BD on the slave side) - (A route evaluation value AD on the master side)" (S63). It is checked whether the difference calculated here is larger than the switching determination threshold value S (S64), and if it is larger, the process proceeds to S65, whereas if it is not larger, the process is ended.

As the threshold value S, for example, if it is assumed that both route evaluation values AD and BD are equivalent, a value of "0" can be used. In this case, if the A route evaluation value AD and the B route evaluation value BD are both the minimum value "0", the threshold is not exceeded and the process is terminated without switching routes, while the B route evaluation value is the minimum value "0". ”, the threshold value is exceeded and the process proceeds to S65.

S65: VLAN-related data of route A on the master side (VLAN information D8, master collection data D6, master collection completion flag F3, etc.) and VLAN-related data of route B on the slave side (VLAN information D8, slave collection data D7, slave collection data D7, etc.) (collection completion flag F4, etc.) is replaced, and then the process ends.

(5) BC conversion transmission processing The details of the BC conversion transmission processing in S28 and S39 will be explained based on FIG. 10.

S71: When the process starts, target PTP ports Po are selected in order to transmit PTP packets.

S72: If the selected target PTP port Po is in a normal state where port transmission is possible, the process advances to S73, whereas if it is not in a normal state, the process advances to S77.

S73, S74: Refer to port BC conversion necessity information D1 to confirm whether BC conversion of the selected target PTP port Po is required. As a result of the confirmation, if necessary, the process proceeds to S75, while if not necessary, conventional PTP processing is executed (S74), and the process proceeds to S76.

S75, S76: Generate a PTP packet changed from original multicast to broadcast (S75). The generated PTP packet is transmitted from the selected target PTP port Po (S76).

S77: It is checked whether the processes of S71 to S76 have been completed for all of the selected target ports Po. As a result of the confirmation, if it has not been completed, the process returns to S71 and the process is restarted, while if it has been completed, the BC conversion transmission process is ended.
(6) Effects According to the redundant route management method described above, the following effects A to C can be obtained.

A: In other words, in a redundant network configuration using VLAN duplication, route evaluation conventionally took time. On the other hand, according to the redundant path management method, if the "Meanpath" value during time synchronization processing becomes negative a certain number of times, a TS abnormality is detected.

Therefore, it is possible to detect a TS abnormality in the main VLAN on the master side at an early stage, quickly switch to the redundant sub-VLAN route on the slave side, and take appropriate measures in this regard. However, since abnormality detection is performed not only on the main VLAN side but also on the sub-VLAN side (S63), switching is not performed if the evaluation on the sub-VLAN side is low.

B: Furthermore, if there is a hub that cannot support VLAN duplication in a redundant network using VLAN duplication, regular multicast transmission can be changed to broadcast transmission by the BC conversion transmission processing in S28 and S39. As a result, hubs that cannot support VLAN duplication are ignored, and PTP processing between PTP hubs 1 and 2 becomes possible.

C: Furthermore, the BC conversion transmission process can be applied to systems other than redundant network configurations based on VLAN duplication, and can be used as a workaround when there is a PTP hub that you want to exclude because it is not subject to management. Note that the BC conversion transmission process can be used without problems even in a network configuration in which hubs compatible with VLAN duplexing and hubs not compatible with VLAN duplexing coexist.

≪Example 1≫
A first embodiment of the redundant route management method will be described based on FIG. 2. Here, a problem (change in PTP processing, functional abnormality/failure, etc.) occurs in hub A in the VLAN duplex connection network configuration shown in Figure 1, and the "Meanpath" value on the PTP hub 2 side becomes negative, causing a TS error. Let us assume a situation in which PTP time synchronization cannot be calculated correctly.

At this time, the PTP hub 2 detects a situation where the "Meanpath" value becomes negative as a TS abnormality (S41), and adds it to the abnormality counter D4 (S42). If this situation continues for a certain period of time (time is variable), the value of the abnormality counter D4 exceeds the TS abnormality determination number D2 (S43).

Therefore, the switching lock flag F1 for the A route is set (S44), and the A route evaluation value AD is forcibly lowered to the lowest value "0" (S61). As a result, the A route and the B route are switched on the condition that S62 to S64 are followed.

In other words, "V2-M-Root" (A root) and "V5-S-Broot" (B root) are replaced by "V5-M-Broot" (B root) and "V2-S-Aroot" (A root). Switch to

At this time, the PTP information packet is used to notify the outside that a TS abnormality has occurred in route A (S45, S46). After that, route A continues to be evaluated by PTP hub 2, but even if hub A recovers from a failure, it cannot return to the master side unless the switching conditions (S61 to S64) are met, such as when route B is disconnected. Can not. Note that as a measure after hub A returns, it is also possible to restart only PTP hub 1 or PTP hub 2 by remote control.

≪Example 2≫
A second embodiment of the redundant route management method will be described based on FIG. 3. Here, the PTP hubs 1-1 and 1-2 are made redundant with a main VLAN (V2-M-Aroot) on the master side and a sub-VLAN (V5-S-Broot) on the slave side.

Furthermore, the PTP hubs 2-1 and 2-2 are made redundant with a main VLAN (V3-M-Root) on the master side and a sub-VLAN (V6-S-Broot) on the slave side.

At this time, communication between PTP hubs 1-1 and 1-2 is via hubs A and B, while communication between PTP hubs 2-1 and 2-2 is via hubs C and D.

If the hubs A to D cannot support VLAN duplication, one or both of the PTP hubs 2-1 and 2-2 and the PTP hubs 2-1 and 2-2 cannot be connected.

In such a case, if the ports of PTP hubs 1-1 to 2-2 that connect to hubs A to D are set to require BC conversion (S73), the data between hubs A to D is transmitted by broadcast, so PTP processing is not performed. It will be transferred without any error. This allows the PTP hub to transmit data between the PTP hubs 2-1 and 2-2 and between the PTP hubs 2-1 and 2-2.

≪Example 3≫
A third embodiment of the redundant route management method will be described based on FIG. 4. Here, a network configuration is assumed in which a PTP hub 1 in area A is connected via PTP to PTP hubs 2 to 7 belonging to areas B to E. At this time, although PTP hubs are used inside areas A, C, and D, hubs that cannot support redundancy through VLAN duplication are used in areas B and E.

According to this network configuration, normally PTP connections are only possible between areas A and C, but the ports of PTP hubs 2 and 4 that pass through area B and the ports of PTP hubs 5 and 7 that pass through area E are converted to BC. If yes (S73), areas A, C, and D are transmitted using regular multicast, while areas B and E are transmitted using broadcast.

Therefore, PTP communication with PTP hub 1 in area A as the master is possible with PTP hub 4 in area B, PTP hub 5 in area C, PTP hub 6 in area D, and PTP hub 7 in area E. .

1 to 7, 1-1, 1-2, 2-1, 2-2...PTP hub A to D...Hub

Claims (7)

A network management method with redundant master-slave routes between time-synchronized masters and slaves,
"Meanpath" is calculated as an average value of the transmission time of the outgoing path and the incoming path between the master and the slave, and if the calculated "Meanpath" continues to be in a negative state, it is determined that the time stamp is abnormal;
A redundant route management method characterized in that the master/slave of the route is switched depending on the determination. 2. The redundant route management method according to claim 1, wherein the time stamp is determined to be abnormal when the number of times the "Meanpath" becomes negative exceeds a preset threshold. 3. The redundant route management method according to claim 1, wherein the determination is made not only for the main route between the master and the slave but also for the secondary route. 4. The redundant route management system according to claim 3, wherein if the subordinate route continues to be in an abnormal state, switching to the main route is not exceptionally performed . 5. The redundant route management method according to claim 1, wherein when the abnormality is determined, an external party is notified. 6. The redundant route management method according to claim 1, wherein packets necessary for the time synchronization are transmitted between the master and slave by broadcast. A method for managing a network with redundant master-slave routes between time-synchronized masters and slaves, the method comprising:
a step of calculating "Meanpath" as an average value of the transmission time of the outbound path and the return path between the master and the slave;
determining that the timestamp is abnormal if the calculated "Meanpath" continues to be in a negative state;
switching the master and slave of the route according to the determination;
A method for managing a redundant route, comprising:

Images