JP2005286720A - Network time synchronizing method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a network time synchronizing method in which a tree form seen from a root is considered to shorten a synchronizing focusing time. <P>SOLUTION: The network time synchronizing method includes a step of synchronizing the clock of a slave with the synchronizing clock of a master in which the function of a communication unit of opposed one node is the master and the function of the communication unit of the other node is as the slave, and a step of synchronizing the clock of the master with the clock of the slave in the node that the master and the slaver are contained. The network time synchronizing method further includes a first step of acquiring the number of subordinate clock synchronizing under the master and the slave by the master and the slave, and a second step of synchronizing the clock of the master with the clock of the slave having the maximum number of the subordinate clock for the node containing the master and the slave. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a time synchronization method for a network in which nodes are distributed.

  As a time synchronization method in a mesh communication network, there is a technique for establishing time synchronization between nodes in an autonomous and distributed manner without requiring a management node (see, for example, Patent Document 1). The mesh communication network is a network in which a plurality of nodes N are distributed and arranged in a mesh shape, each node includes at least one communication device, and opposing nodes establish a link between the communication devices. In such a technique, each node establishes synchronization only with an opposite node, and a plurality of nodes are sequentially configured in a tree shape as a whole. Therefore, even when there is a partial node failure, it is possible to maintain the tree within the range connected by the link.

  For example, the technology based on Patent Document 1 synchronizes the slave clock with the master synchronization clock, with the function of the communication device of one opposite node as the master and the function of the communication device of the other node as the slave. The relationship between the master and slave facing each other via the link can be switched by negotiation between the two. The node in which the master and the slave are accommodated synchronizes the master clock with the slave clock.

  Among the plurality of masters connected to the synchronized link in this way, a master functioning as a clock supply source is defined as a root, and a plurality of nodes are configured in a tree shape. Root R is assumed to exist only in a tree in which synchronization is established.

  In the mesh communication network, not all nodes are synchronized, and a plurality of time synchronization trees exist. Even if a plurality of communication devices exist in one node, not all clocks are always synchronized. In this case, one node belongs to a different tree.

Japanese Patent Laid-Open No. 2003-274349

  However, according to the time synchronization method of Patent Document 1, it is determined which clock to synchronize using only the priority set in advance for the clock of each node. Therefore, the tree shape seen from the root cannot be taken into account, so that there is a case where the tree is biased. Such a tree has an increased convergence time and is inefficient in realizing dependent synchronization from the root.

  Accordingly, an object of the present invention is to provide a network time synchronization method that takes into consideration the shape of a tree viewed from the root in order to shorten the synchronization convergence time.

  A time synchronization method in a network targeted by the present invention is a network in which a plurality of nodes are distributed and each node includes at least one communication device, and an opposite node establishes a link between the communication devices. The function of the communication device of one of the opposing nodes is a master, the function of the communication device of the other node is a slave, the step of synchronizing the slave clock to the master synchronization clock, and the master and slave are accommodated The node has a step of synchronizing the master clock to the slave clock, and among the plurality of masters connected to the synchronized link, the master functioning as a clock supply source is defined as a root, and the plurality of nodes are defined. In a network configured in a tree shape.

The network time synchronization method according to the present invention includes:
A first step in which the master and the slave acquire the number of clocks under the route synchronized under the route;
The node in which the master and the slave are accommodated includes a second step of synchronizing the master clock to the slave clock having the maximum number of clocks under the root.

According to another embodiment of the time synchronization method of the present invention, the second step is:
A slave other than the slave having the maximum number of clocks under the root route exchanges the function with the master facing the slave, thereby changing the slave to the master, and changing the master to the slave; and
Preferably, the method further comprises the step of synchronizing the changed master clock with the slave clock having the maximum number of clocks under the root.

Further, according to another embodiment of the time synchronization method of the present invention, when the master is not accommodated in the node, the first step defines the slave accommodated in the node as a leaf,
The leaf transmits a hop count notification message with a subordinate clock count of 1 to the upper node that is the master of the leaf;
The master that has received the hop count notification message holds the total number of subordinate clocks included in the message, and transmits a hop count notification message including the total subordinate clock counts to the upper node that is the master, and hop count notification The route that received the message holds the total number of subordinate clocks included in the message, and transmits a route ID notification message including the total number of subordinate clocks to the lower node;
It is also preferable that the master and the slave that have received the route ID notification message further include a step of holding the number of clocks under the route included in the message.

Furthermore, according to another embodiment of the time synchronization method of the present invention,
For the first step, the master and slave each further acquire the number of route hops from the node to the route,
Regarding the second step, when the number of clocks under the route of the master and slave is the same, it is also preferable to synchronize the master clock with the clock of the slave having the minimum number of route hops.

Furthermore, according to another embodiment of the time synchronization method of the present invention,
For the second step, for a slave other than the slave having the minimum number of root hops and having a number of root hops greater than or equal to the minimum number of root hops by exchanging functions with a master facing the slave Changing the slave to a master and changing the master to a slave;
Preferably, the method further comprises the step of synchronizing the changed master clock to the slave clock having the minimum number of route hops.

Furthermore, according to another embodiment of the time synchronization method of the present invention,
A route transmitting a route ID notification message in which the number of route hops is 0 to a lower node that is a slave of the route;
The slave that has received the route ID notification message further has a step of holding the number of route hops included in the message and transmitting a route ID notification message including the number of route hops incremented by one to a lower-level node that is a slave. Is also preferable.

Also, the network time synchronization method in the present invention is:
A first step in which each of the master and the slave obtains the number of route hops from the node to the route;
A second step of synchronizing a master clock having a number of root hops greater than or equal to the minimum number of root hops with a clock of a slave having a minimum number of root hops for a node in which the master and slave are accommodated; Features.

According to another embodiment of the time synchronization method of the present invention,
The second step is
For a slave other than the slave having the minimum number of root hops, and having a number of root hops that is two or more greater than the minimum number of root hops, the slave is changed to the master by exchanging the function with the master facing the slave. And changing the master to the slave,
Preferably, the method further comprises the step of synchronizing the changed master clock to the slave clock having the minimum number of route hops.

Furthermore, according to another embodiment of the time synchronization method of the present invention,
The first step is
A route transmitting a route ID notification message in which the number of route hops is 0 to a lower node that is a slave of the route;
The slave that has received the route ID notification message further has a step of holding the number of route hops included in the message and transmitting a route ID notification message including the number of route hops incremented by one to a lower-level node that is a slave. Is also preferable.

Furthermore, the network time synchronization method according to the present invention includes:
A root is obtained for each slave that synchronizes via a link with a node that accommodates the route, a first step of obtaining the number of leaf hops to a leaf that is a slave located at the end of the tree;
When there is a slave having two or more leaf hops more than the minimum number of leaf hops among a plurality of slaves connected to the node in which the route is accommodated, the route functions as a slave having the minimum number of leaf hops. A second step of changing the route to a slave and changing what was a slave to a route by exchanging
And a third step of synchronizing the slave clock with the root clock.

  According to another embodiment of the time synchronization method of the present invention, when the master is further accommodated in the node where the route is accommodated, the master clock is synchronized with the clock of the slave whose function is exchanged. It is also preferable to further include a step.

According to another embodiment of the time synchronization method of the present invention, the first step is:
A leaf transmits a leaf hop count notification message with a leaf hop count of 1 to an upper node serving as a master of the leaf;
The master that has received the leaf hop count notification message further has a step of holding the leaf hop count included in the message and transmitting the leaf hop count notification message including the leaf hop count incremented by 1 to the master upper node. It is also preferable.

  According to the network time synchronization method of the present invention, the shape of the tree seen from the root can be considered in order to shorten the synchronization convergence time. Since it can be subordinate to the clock of the tree with a large number of subordinate clocks, and can be subordinate to the clock with a small number of hops to the root, jitter can be suppressed and accurate synchronization timing can be realized. . Furthermore, the number of hops in the entire tree can be shortened by moving the route to the optimum position in the tree. According to the present invention, no specific management node is required, and each node operates in an autonomous distributed manner for time synchronization.

  In particular, since the synchronous network can be configured or maintained in an autonomous and distributed manner in response to changes in the network topology, it is effective for use in a nomadic / ad hoc network that assumes movement of radio stations. In addition, the tree construction technique in the present invention is not limited to wireless, but can also be applied to a wired network such as the Internet. For example, even in a time synchronization method such as NTP (Network Time Protocol), it is possible to configure an optimal tree and maintain synchronization without setting a hierarchical relationship between reference nodes and nodes in advance. It is extremely effective.

  Hereinafter, the best embodiment of the present invention will be described in detail with reference to the drawings.

The “clock” refers to a communication device that terminates via a link, and includes a root clock R, a master clock M, a slave clock S, and a leaf clock L.
“Subordinate clock number” Nall means the total number of clocks subordinate to the own clock. When a plurality of clocks are subordinate to the own clock, the subordinate clock number is obtained by adding the subordinate clock numbers of the subordinate clocks.
The “number of clocks under the root” Rall means the total number of clocks subordinate to the root. Each clock holds and serves as a reference for clock synchronization.
The “number of route hops” Rhop is the number of hops indicating the distance from the route to the own node when the route is 0.
The “number of leaf hops” Lhop refers to the number of hops indicating the distance from the leaf to the own node when the leaf that is the slave at the end of the tree is 0. When there are a plurality of leaves from the own node, the maximum number of hops is set as the number of leaf hops of the own node.

  The method of the present invention for constructing a new tree is realized by having each clock hold the “number of clocks under the root” Rall and the “number of root hops” Rhop. By considering the “number of clocks under the route” Rall, the node can be subordinate to a clock with a large number of clocks under the route, and the time until convergence establishment in the entire network can be shortened. Also, by considering the “maximum number of hops” Rhop, it is possible to subordinate to a tree that minimizes the number of hops to the leaf and optimize the tree structure, ie, from the reference node of the tree to each node The maximum number of hops can be minimized.

  The method of the present invention for optimizing an already constructed tree includes a “hop number shortening method” and a “route moving method”. The "hop count shortening method" is to switch the dependent destination by reversing the dependency relationship of the clock with a large number of hops when the difference in the number of hops from the route is more than a certain value among multiple clocks in the local node. The number of hops of lower clocks including its own node is shortened, and the maximum number of hops in the tree is shortened. The “route moving method” is a method in which, in a tree once constructed, the route is moved to an optimal position as a whole without shortening the tree structure, and the maximum number of hops in the tree is shortened. This moves the route to an optimum position by a series of two operations of “root slave reverse rotation” and “route master reverse rotation”.

  In order to realize these methods, a “route ID notification message” and a “hop number notification message” are used.

  The “root ID notification message” Rmes includes a tree ID, a root hop count Rhop, and a root subordinate clock count Rall. This message notifies a lower node that the route state has changed (for example, link disconnection or the like). The clock that has received this message holds these values (tree ID, Rhop, and Rall), increments the number of root hops Rhop by 1, and transmits a route ID notification message to the lower clock.

  The “hop count notification message” Hmes includes a leaf hop count Lhop and a subordinate clock count Nall. This message notifies the upper node that the number of hops has changed due to link independence or tree ID change. The clock that has received this message retains these values (Lhop and Nall), increments the leaf hop number Lhop by 1, re-adds the subordinate clock number, and transmits a hop number notification message to the upper clock.

  FIG. 1 is a tree configuration diagram showing message transmission and reception in the present invention.

  According to FIG. 1, the master M0 is synchronized with the route R in the node N0. The slave S1 of the node N10 is synchronized with the route R of the upper node N0, and the slave S1 of the node N11 is synchronized with the master M0 of the upper node N0. For the node N10, the three masters M1 are synchronized with the slave S1. The master M1 is synchronized with the slave S2 of the lower nodes N21 to N23 of the node N10.

  Each node clock (master M, slave S, route R, and leaf L) has Rhop “number of root hops”, Lhop “number of leaf hops”, Nall “number of subordinate clocks”, and Rall “number of clocks under subordinate route”. And hold.

  The node N0 transmits a route ID notification message Rmes (ID, Rhop, Rall) to the lower nodes N10 and N11. The route hop count Rhop of the route ID notification message Rmes transmitted by the node N0 is zero.

  In response to the route ID notification message Rmes, the slave S1 of the node N10 transmits ACK to the node N0. The slave S1 of the node N0 retains the route hop number 0 incremented by 1 by receiving the route ID notification message Rmes. Further, each master M1 synchronized with the slave S1 similarly holds the route hop count Rhop = 1, and further transmits a route ID notification message Rmes with the root hop count Rhop = 1 to the lower nodes N21 to N23. To do. On the other hand, the node N10 receives ACK from the lower nodes N21 to N23.

  Thus, the route ID notification message Rmes transmitted from the route R reaches the leaf while being relayed from the upper level to the lower level. In response to this, the leaf transmits a hop number notification message Hmes (Lhop, Nall) to the upper node. In the hop count notification message Hmes transmitted by the leaf, the leaf hop count Lhop is 0 and the subordinate clock count Nall is 1.

  The master M1 of the node N10 receives the hop count notification message Hmes (Lhop = 5, Nall = 10) from the node N21. At this time, the master M1 retains the leaf hop number incremented by 1 and retains the subordinate clock count of 10 incremented by one. Furthermore, the slave S1 with which the master M1 is synchronized holds the maximum number of leaf hops among all the masters M1. Here, among the three masters M1, the maximum number of leaf hops is seven. In addition, the slave S1 sets a subordinate clock number 40 that is the sum of the total number of subordinate clocks (11 + 10 + 15 = 36) and the number of clocks in the node N10 (slave 1 + master 3 = 4) for all masters M1. Hold.

  The master M0 of the node N0 has received the hop count notification message Hmes (Lhop = 5, Nall = 8) from the node N11. As a result, the master M0 holds the leaf hop number 6 which is incremented by 1 to the received leaf hop number 5. Further, the master M0 holds the subordinate clock number 9 which is the sum of the received subclock number 8 and its own clock number 1.

  Finally, the route R of the node N0 receives the hop count notification message Hmes (Lhop = 7, Nall = 40) from the slave S1 of the node N10. As a result, the route R retains the leaf hop count 8 incremented by 1 to the maximum value (7> 6) of the received leaf hop count from the relationship with the master M0. In addition, the route R holds a subordinate clock number 50 that is the sum of the total number (40 + 9) of the received subordinate clock numbers and its own clock number 1.

  As described above, each clock can acquire a predetermined value by transmitting and receiving the route ID notification message Rmes and the hop number notification message Hmes.

  FIG. 2 is a tree configuration diagram when a link down occurs in the configuration of FIG.

  Assume that a link down occurs in the subordinate tree of the node N23, the number of clocks Nall subordinate to N23 decreases to 8, and the number of leaf hops Lhop decreases to 3. At this time, the node N23 transmits a hop count notification message Hmes (Lhop = 3, Nall = 8) to the upper node N10.

  The master M1 of the node N10 that has received the hop count notification message Hmes holds the leaf hop count 3 incremented by 1, and holds the subordinate clock count 8. Further, in the slave S1 with which the master M1 is synchronized, the maximum number of leaf hops among the three masters M1 is six. Therefore, the slave S1 has a subordinate clock number 33 which is the sum of the total number of subordinate clocks (11 + 10 + 8 = 29) and the clock number 4 (slave 1 + master 3 = 4) in the node N10 for all masters M1. Hold.

  The slave S1 of the node N10 transmits a hop count notification message Hmes (Lhop = 6, Nall = 33) to the route R of the upper node N0. As a result, the route R holds the leaf hop count 7 incremented by 1 to the maximum value 6 of the received leaf hop count from the relationship with the master M0. In addition, the route R holds a subordinate clock number 43 that is the sum of the total value (33 + 9) of the received subordinate clock numbers and its own clock number one.

  As described above, even when a change occurs in the tree, each clock can acquire a predetermined value by transmitting and receiving the route ID notification message Rmes and the hop number notification message Hmes.

  FIG. 3 is a flowchart when the node receives the route ID notification message Rmes.

(S301) The host clock (slave) of the node receiving the route ID notification message Rmes (ID, Rhop, Rall) updates the tree ID, the number of route hops Rhop, and the number of clocks under the route Rall.
(S302) The upper clock transmits ACK to the transmission source node of the route ID notification message.
(S303) It is determined whether or not there is a master of another link subordinate to the upper clock in the own node. That is, it is determined whether or not the own node is a leaf node.
(S304) If the own node is not a leaf node, the processing of S305 to S307 is executed for all masters subordinate to the higher clock.
(S305) The lower clock (master) updates the tree ID, the number of root hops Rhop, and the number of clocks subordinate to the root Rall.
(S306) A route ID notification message in which the number of route hops is incremented by 1 is transmitted to a lower node synchronized with the lower clock (master).
(S307) ACK is received from the lower node.
(S308) The processes of S305 to S307 are repeated for all masters subordinate to the higher clock. Thereafter, the process ends.
(S309) If the local node is a leaf node, it is determined whether or not this route ID notification message is a response to the hop number notification message. This determination is made based on a flag included in the message. If it is a response to the hop number notification message, the process is terminated.
(S310) If the route ID notification message is not a response to the hop number notification message, the hop number notification message Hmes (Lhop = 0, Nall = 1) is transmitted to the upper node and processed. Exit.

  FIG. 4 is a flowchart when the node receives the hop count notification message Hmes.

(S401) The lower clock (master or route) of the own node that has received the hop count notification message Hmes (Lhop, Nall) is incremented by 1 to the notified leaf hop count Lhop as the number of leaf hops held by the lower clock Holds the value. Further, the notified subordinate clock number Nall is held as the subordinate clock number held by the lower clock. These past values are also temporarily held in order to determine a change from the current value.
(S402) With respect to the number of leaf hops and the number of subordinate clocks, it is determined whether or not a change has occurred between a past value and a current value. If there is no change, the process is terminated.
(S403) If there is a change in these values, it is determined whether there is a clock subordinate to the node. That is, it is determined whether or not the node is the root.
(S404) If the own node is not the root, the upper clock (slave) of the own node holds the maximum value of the number of leaf hops of the currently subordinate clock as the number of leaf hops held by the upper clock. Also, as the number of subordinate clocks held by the upper clock, the sum of the total number of subordinate clocks of the currently subordinate lower clock and the number of subordinate clocks currently subordinate is held. These past values are also temporarily held in order to determine a change from the current value.
(S405) It is determined whether there is an upper node with which the upper clock is synchronized.
(S406) If the own node is a route in S403 or if an upper node exists in S405, it is determined whether or not this hop count notification message is a response to the route ID notification message. To do. This determination is made based on a flag included in the message. If it is a response to the route ID notification message, the process is terminated.
(S407) If the hop count notification message is not a response to the route ID notification message, the route ID notification message Rmes (ID, Rhop) is transmitted to the lower node, and the process is terminated.
(S408) If there is no upper node in S405, it is determined whether or not a change has occurred in the past value and the current value for the number of leaf hops and the number of subordinate clocks. If there is no change, the process is terminated.
(S409) If there is a change in the number of leaf hops and the number of subordinate clocks, the number of leaf hops is incremented by 1 for the upper node with which the slave clock is synchronized, and then the hop number notification message Hmes (Lhop , Nall) is transmitted to finish the process.

  FIG. 5 is an explanatory diagram showing the relationship between the route ID notification message and the hop number notification message.

  According to FIG. 5, when the dependency between the master and the slave is resolved in the node, the master functions as a root. At this time, the slave transmits a hop number notification message to the upper node, and the route transmits a route ID notification message to the lower node. The leaf node that has received the route ID notification message transmits a hop number notification message to the upper node in response thereto. In addition, the route that has received the hop number notification message transmits a route ID notification message to the lower node in response thereto.

  The route ID notification message and the hop number notification message are always in a one-to-one relationship. When a route ID notification message is transmitted, a hop number notification message is transmitted, and when a hop number notification message is transmitted, a route ID notification message is transmitted. Both messages include a flag for identifying whether the currently transmitted message is based on a route ID notification message or a hop number notification message. This prevents the message from being transmitted without reproduction.

  FIG. 6 is a flowchart of clock synchronization processing in the present invention.

(S601) It is determined whether or not there is an unsynchronized clock in the own node. If all clocks are synchronized, the process proceeds to S605.
(S602) If there is an unsynchronized clock, it is determined whether or not the number of clocks Rall under the root of all the clocks is the same.
(S603) If the root subordinate clock numbers Rall of all the clocks are not the same, the slave having the largest root subordinate clock number is defined as the reference clock. Thereafter, a clock that is not synchronized in the node is set as a synchronization target.
(S604) If the number of clocks Rall under the root of all clocks is the same, the clock having the smallest number of route hops is defined as the reference clock. Thereafter, a clock that is not synchronized in the node is set as a synchronization target.
(S605) If all clocks are synchronized in the own node, it is determined whether or not there is a clock having a root hop count Rhop of 2 hops or more in the own node. If there is no clock that is more than 2 hops away, the process ends.
(S606) If there is a clock that is two or more hops away, the slave with the smallest number of route hops is defined as the reference clock. Thereafter, a clock that is two hops away from the reference clock is set as a synchronization target.
(S607) The processing of S608 to S611 is executed for all clocks other than the reference clock.
(S608) It is determined whether or not the clock is a slave. If it is not a slave, that is, if it is a master, the process proceeds to S610.
(S609) If the clock is a slave, slave / master exchange processing is performed. This method is the same as that disclosed in Patent Document 1, and exchanges the functions of the slave and the master that faces the slave through a link. Thereby, the clock becomes the master.
(S610) The master is synchronized with the reference clock.
(S611) The master synchronized with the reference clock transmits a route ID notification message to the lower node.
(S612) S608 to S611 are repeated for all other clocks.
(S613) The slave serving as the reference clock transmits a hop count notification message to the upper node.

  FIG. 7 is an explanatory diagram of shortening the number of hops only when two or more hops are away.

  As shown in FIG. 7 (1), when the difference from the newly subordinate clock is 1 hop, the number of hops to the leaf node is the same even after switching. As shown in FIG. 7 (2), when the difference from the newly subordinate clock is 2 hops, the number of hops to the leaf node is reduced by 1 (L3-> L2). As shown in FIG. 7 (3), when the difference from the newly subordinate clock is 3 hops, the number of hops to the leaf node is reduced by 2 (L3-> L1). As described above, when the difference from the newly subordinate clock is 2 hops or more, there is an effect of shortening the number of hops. The operation for shortening the number of hops naturally occurs not only for the leaf but also for the slave to which the master is subordinate.

  FIG. 8 is an explanatory diagram of the hop number shortening operation in the node.

(S81) The slave S1 of the node N3 is synchronized with the route R of the node N1, and the master M1 is synchronized with the slave S1. The slave S3 synchronizes with the master M1 through the same tree. Further, the master M3 is synchronized with the slave S3, and the master M3 is synchronized with the leaf L4 of the node N2.
(S82) At this time, the slave S1 having the smallest number of hops becomes the reference clock. First, the dependency between the slave S3 and the master M1 is eliminated. Next, the master M1 from which the subordinate has been removed subordinates to the slave S1 of the reference clock.
(S83) Next, the slave S3 whose dependency has been resolved exchanges the function with the master M2 that is opposed through the link. The master M1 exchanged from the slave is subordinate to the slave S1 of the reference clock.

  FIG. 9 is a flowchart of route movement in the present invention.

(S901) It is determined whether the own node accommodates the route. If the route is not accommodated, the process ends.
(S902) It is determined whether or not the route forms a tree with two or more links.
(S903) When there is only one link in the route, it is determined whether or not the number of leaf hops of the route is 2 or more. If it is not 2 or more, the process is terminated.
(S904) If the number of leaf hops of the route is two or more, the link destination of the route is set as a route movement target.
(S905) When there are two or more links in the route, it is determined whether or not the difference between the number of the top two leaf hops with the largest number of leaf hops is 2 or more with respect to the number of leaf hops for each link. If it is not 2 or more, the process is terminated.
(S906) If the difference in the number of leaf hops is 2 or more, the link destination having the maximum number of leaf hops is set as a route movement target.
(S907) It is determined whether the master is accommodated in the link destination node to be moved. If the master is not accommodated, the process ends.
(S908) If the master is accommodated in the link destination node, the route and the master are reversed.

  FIG. 10 is an explanatory diagram of route movement.

  As shown in FIG. 10, the number of hops in the entire tree must be increased as the route approaches the end of the tree. Therefore, by repeating the route / slave reversal and the route / master reversal, the number of hops seen from the route can be averaged by moving the route to the center of the tree.

  For a node having a route, route movement is performed when there is a difference of 2 or more in the maximum number of hops from each link destination. With respect to FIG. 10, in state (1), the route is 4 hops to the right and 0 hops to the left, so movement is performed. In state (2), the root / slave is reversed. Even in state (3), since the route is 3 hops to the right and 1 hop to the left, there is a difference of 2 or more, so the route moves further. In state (4), there are 2 hops to the right and 2 hops to the left, and there is no difference in hops, so no route movement is performed.

  In the route movement operation, the route does not move to an optimal position in one operation, but moves up to an adjacent node in one operation. After moving to an adjacent node, a route movement operation occurs again. If it is determined that movement is necessary as a result of the hop count calculation, the route is moved again.

  FIG. 11 is an explanatory diagram of route / master reverse rotation.

  It is assumed that a route and a master are accommodated in the node N1. At this time, the master is temporarily set as a root, and the synchronization tree on the right side is made independent. Then, the left route is set as the master. Since such an operation is completed instantaneously, it can be executed without an independent / subordinate resolution operation. Since the synchronization timings of the two routes in the node N1 are considered to be substantially the same, this operation can be completed in a shorter period of time than the clock synchronization process.

  FIG. 12 is an explanatory diagram of the effects of hop count reduction and route movement according to the present invention.

  FIG. 12A is a tree configuration diagram that can improve the number of hops. According to this, the maximum number of hops is 3. FIG. 12B is a tree configuration diagram when the number of hops is shortened from FIG. According to this, the maximum number of hops is 2. FIG. 12C is a tree configuration diagram when the route is moved from FIG. This also sets the longest hop count to two. If the difference in the number of hops is 1 as shown in FIG. 12D, there is no difference in effect regardless of whether the hop number is shortened or the route is moved.

  According to the above-described various embodiments of the network time synchronization method of the present invention, those skilled in the art can easily make various changes, modifications, and omissions in the technical idea and scope of the present invention. The above description is merely an example, and is not intended to be restrictive. The invention is limited only as defined in the following claims and the equivalents thereto.

It is a tree block diagram showing transmission / reception of the message in this invention. FIG. 2 is a tree configuration diagram when a link down occurs in the configuration of FIG. 1. It is a flowchart in case a node receives route ID notification message Rmes. It is a flowchart in case a node receives the hop number notification message Hmes. It is explanatory drawing which shows the relationship between a route ID notification message and a hop number notification message. 5 is a flowchart of clock synchronization processing in the present invention. It is explanatory drawing of shortening the number of hops only when it leaves | separates 2 hops or more. It is explanatory drawing of the hop number shortening operation | movement within a node. It is a flowchart of route movement in the present invention. It is explanatory drawing of route movement. It is explanatory drawing of route / master reverse rotation. It is explanatory drawing of the effect of the hop number shortening and route movement of this invention.

Claims (12)

In a network in which a plurality of nodes are distributed, each node includes at least one communication device, and opposing nodes establish a link between the respective communication devices,
The function of the communication device of one opposite node as a master, the function of the communication device of the other node as a slave, the step of synchronizing the clock of the slave with the synchronization clock of the master,
The node in which the master and the slave are accommodated has a step of synchronizing the clock of the master with the clock of the slave, and functions as a clock supply source among a plurality of masters connected to the synchronized link In the time synchronization method in the network in which the master to be defined as a root and the plurality of nodes are configured in a tree shape,
A first step in which the master and the slave acquire the number of clocks under the route synchronized under the route;
A network time synchronization method comprising: a second step of synchronizing a clock of the master with a clock of a slave having a maximum number of clocks under the root of the node in which the master and the slave are accommodated. The second step includes
A slave other than the slave having the maximum number of clocks under the root route exchanges the function with the master facing the slave, thereby changing the slave to a master, and changing the master to a slave; and
The network time synchronization method according to claim 1, further comprising the step of synchronizing the changed master clock with the slave clock having the maximum number of clocks under the maximum route. The first step includes
When the master is not accommodated in the node, the slave accommodated in the node is defined as a leaf,
The leaf sends a hop count notification message with the subordinate clock count to 1 to the upper node that is the master of the leaf;
The master that has received the hop number notification message holds the total number of subordinate clocks included in the message, and transmits a hop number notification message that includes the total subordinate clock number to the upper node serving as a master;
The route that has received the hop number notification message holds the total number of subordinate clocks included in the message, and transmits a route ID notification message including the total number of subordinate clocks to the lower node.
The network time synchronization method according to claim 1, further comprising: a step in which the master and the slave that have received the route ID notification message hold the number of clocks under the route included in the message. For the first step, the master and the slave each further acquire the number of route hops from the node to the route,
In the second step, when the number of clocks under the route of the master and the slave is the same, the clock of the master is synchronized with the clock of the slave having the minimum number of route hops. The time synchronization method for a network according to any one of the above. For the second step, for a slave other than the slave having the minimum number of root hops and having a number of root hops greater than or equal to the minimum number of root hops, the function is exchanged with the master facing the slave. To change the slave to the master, change the master to the slave, and
5. The network time synchronization method according to claim 4, further comprising the step of synchronizing the changed master clock with the slave clock having the minimum number of route hops. Sending a route ID notification message in which the route has a route hop count of 0 to a subordinate node serving as a slave of the route;
The slave that has received the route ID notification message holds the number of route hops included in the message, and further transmits a route ID notification message including the number of route hops incremented by 1 to a lower-level node that is a slave. The network time synchronization method according to claim 4, wherein the network time synchronization method is provided. In a network in which a plurality of nodes are distributed, each node includes at least one communication device, and opposing nodes establish a link between the respective communication devices,
The function of the communication device of one opposite node as a master, the function of the communication device of the other node as a slave, the step of synchronizing the clock of the slave to the synchronization clock of the master,
The node in which the master and the slave are accommodated has a step of synchronizing the clock of the master with the clock of the slave, and functions as a clock supply source among a plurality of masters connected to the synchronized link In the time synchronization method in the network in which the master to be defined as a root and the plurality of nodes are configured in a tree shape,
A first step in which each of the master and the slave obtains the number of route hops from the node to the route;
A second step of synchronizing a clock of a master having a number of root hops greater than or equal to the minimum number of root hops by a clock of a slave having a minimum number of root hops with respect to a node accommodating the master and the slave; A time synchronization method for a network, comprising: The second step includes
For a slave other than the slave having the minimum number of root hops, the slave having a number of root hops that is two or more larger than the minimum number of root hops, by exchanging functions with the master facing the slave, Changing to master and changing master to slave,
8. The network time synchronization method according to claim 7, further comprising the step of synchronizing the changed master clock with the slave clock having the minimum number of route hops. The first step includes
Sending a route ID notification message in which the route has a route hop count of 0 to a subordinate node serving as a slave of the route;
The slave that has received the route ID notification message holds the number of route hops included in the message, and further transmits a route ID notification message including the number of route hops incremented by 1 to a lower-level node that is a slave. The network time synchronization method according to claim 8, wherein the network time synchronization method is provided. In a network in which a plurality of nodes are distributed, each node includes at least one communication device, and opposing nodes establish a link between the respective communication devices,
The function of the communication device of one opposite node as a master, the function of the communication device of the other node as a slave, the step of synchronizing the clock of the slave to the synchronization clock of the master,
The node in which the master and the slave are accommodated has a step of synchronizing the clock of the master with the clock of the slave, and functions as a clock supply source among a plurality of masters connected to the synchronized link In the time synchronization method in the network in which the master to be defined as a root and the plurality of nodes are configured in a tree shape,
A first step of obtaining, for each slave synchronized via a link with a node accommodating the route, a leaf hop count to a leaf that is a slave located at the end of the tree;
When there is a slave having a number of leaf hops more than the minimum number of leaf hops among the plurality of slaves connected to the node in which the route is accommodated, the slave having the minimum number of leaf hops By exchanging functions with the second step of changing the route to a slave and changing the slave to a route;
And a third step of synchronizing a clock of the slave with a clock of the route.   The method further comprises a fourth step of synchronizing the clock of the master with the clock of the slave whose function has been exchanged when a master is further accommodated in the node in which the route is accommodated. The network time synchronization method according to claim 10. The first step includes
The leaf sends a leaf hop count notification message with a leaf hop count of 1 to an upper node that is the master of the leaf;
The master that has received the leaf hop count notification message holds the leaf hop count included in the message and transmits a leaf hop count notification message including the leaf hop count incremented by 1 to the master upper node. The network time synchronization method according to claim 10 or 11, further comprising:

Images