JP2008118274A - Network communication apparatus and communication program for ring type network - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To enable management information needed to maintain a ring to be shared in the entire ring. <P>SOLUTION: A plurality of nodes are constituted in a ring shape and a communication program multicasts a message, transmitted from an arbitrary node as an origin, in the ring by transferring the message in the ring. When the message does not reach a node at a transmission destination, a message indicating that the node enters a disconnected state is generated and transmitted forward and backward in the ring except this transmission destination. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a ring network in which a plurality of nodes are configured in a ring shape, and a message transmitted from an arbitrary node is transferred within the ring to multicast a message within the ring.

  For example, as disclosed in Patent Documents 1 and 2, a ring network that multicasts a message by configuring a plurality of nodes in a ring shape and transferring the message unidirectionally in the ring has been conventionally known. ing. Japanese Patent Application Laid-Open No. H10-228688 discloses a countermeasure against a communication path failure in an ad hoc network. Specifically, a certain communication terminal detects a disconnection sign of the communication path on the upstream side of the message, and notifies this to the upstream communication terminal as a pre-disconnection state. Upon receiving this notification, the upstream communication terminal transmits a message to the transmission destination through a route other than the route corresponding to the pre-disconnection state.

Japanese Patent Laid-Open No. 5-219084 JP 2000-4246 A JP 2005-64723 A

  By the way, some ring networks are premised on sharing information in a ring when a node constituting the communication path of the ring is disconnected. In this case, the node that detects the disconnection of a certain node starts from itself and notifies each node of a notification message to that effect. However, when this notification message is transferred unidirectionally within the ring, the following two problems arise. First, there is a problem regarding communication time required for message sharing in the ring. When the number of nodes in the ring is n (n ≧ 3), when messages are sequentially transferred in one direction between adjacent nodes, the number of transfers is n. Assuming that the message arrival time between nodes is the same among all nodes, the communication time is proportional to the number of transfers. Therefore, when application to an application (for example, a real-time competitive network game) that needs to share a message in real time is assumed, this may be an overhead that impairs real-time performance. Second, there is a problem regarding the reliability of the communication path. In message transfer in one direction, if a part of the communication path is disconnected, subsequent message transfer to downstream nodes is delayed. This is a big problem in a mobile terminal (for example, a mobile phone) that is more likely to cause a communication path disconnection than a fixed terminal.

  Accordingly, an object of the present invention is to share management information necessary for maintaining a ring quickly and reliably throughout the ring.

  In order to solve this problem, according to a first aspect of the present invention, a plurality of nodes are configured in a ring shape, and a message transmitted from an arbitrary node is transferred in the ring to thereby generate a message in the ring. A network communication device that is used as a node in a ring network that multicasts. The network communication device includes a communication unit, a first node management unit, an arrival confirmation unit, and a starting point control unit. The communication unit transmits and receives messages to and from other nodes. The first node management unit manages, for each node, a status indicating whether or not the node is connected in the communication path in the ring. The arrival confirmation unit determines whether or not the first message transmitted from its own node has reached the destination node via the communication unit. When the arrival confirmation unit determines that the first message has not reached the destination node, the starting point control unit accesses the node management unit and updates the status regarding this node to a disconnected state. At the same time, the origin control unit transmits a second message indicating that the node has been updated to the disconnected state in the forward direction and the reverse direction in the ring excluding the transmission destination via the communication unit. The status update content is notified to each node in the ring.

  Here, in the first invention, a second node management unit, a transfer determination unit, and a transfer control unit may be further provided. The second node management unit manages the message received by its own node by forwarding within the ring for each node that is the starting point of the message. When the third message transmitted from a certain starting point is received by the communication unit in either the forward direction or the reverse direction, the transfer determination unit accesses the second node management unit to It is determined whether or not the third message has been received for the node that is the starting point of the message. When it is determined that the third message has not been received, the transfer control unit accesses the first node management unit, updates the status regarding the node notified by the third message to a disconnected state, After accessing the second node management unit and updating the identification information regarding the node that is the starting point of the third message, the third message is transmitted in the transfer direction in the ring via the communication unit. Further, when it is determined that the third message has been received, the transfer control unit terminates the transfer of the third message in the transfer direction at its own node without transmitting the third message.

  In the first invention, the second node management unit may manage the identification information for identifying the message transferred in the ring for each node that is the starting point of the message. In this case, the starting point control unit numbers the first identification information to be given to the second message, and transmits the second message to which the first identification information is given in the forward direction and the reverse direction. In addition, the transfer determination unit compares the second identification information extracted from the third message with the identification information about the node that is the starting point of the third message in the node management unit. It is determined whether or not has been received.

  In the first invention, the identification information may include a message ID assigned to each message in node units and a node ID unique to the node. In this case, the starting point control unit assigns a newly assigned message ID and its own node ID indicating that it is the starting point to the second message using the identification information. Further, the transfer determination unit specifies the node that is the starting point of the third message based on the node ID extracted from the third message.

  A second invention is a ring network in which a plurality of nodes are configured in a ring shape, and a message transmitted from an arbitrary node is transferred in the ring to multicast a message in the ring. Provided is a communication program installed in a network communication device used as a node. This ring network communication program causes a network communication device to execute a method including the following steps.

(1) Step 1
A node list is prepared for managing, for each node, a status indicating whether or not a node is connected on a communication path in the ring.
(2) Step 2
It is determined whether or not the first message transmitted from its own node has reached the destination node.
(3) Step 3
When it is determined that the first message has not arrived at the destination node, the status regarding this node in the node list is updated to a disconnected state.
(4) Step 4
By sending a second message indicating that this node has been updated to the disconnected state in the forward and backward directions within the ring, the status update content is notified to each node in the ring except this node. To do.

  Here, in the second invention, the node list may manage the message received by its own node by the transfer in the ring for each node that is the starting point of the message. In this case, 2nd invention makes a network apparatus perform the method which further has the following steps.

(5) Step 5
When a third message transmitted from a certain origin is received in either the forward or reverse transfer direction, the third message has been received for the node that originated the third message in the node list. It is determined whether or not there is.
(6) Step 6
When it is determined that the third message has not been received, the status regarding the node notified by the third message in the node list is updated to the disconnected state, and the node that is the starting point of the third message in the node list The third message is transmitted in the transfer direction within the ring.
(7) Step 7
If it is determined that the third message has been received, the third message transfer in the transfer direction is terminated at its own node without transmitting the third message.

  In the second invention, it is preferable that the node list manages identification information for identifying a message transferred in the ring for each node that is a starting point of the message. In this case, the fourth step is the step of numbering the first identification information to be given to the second message, and the second message to which the first identification information is given are forward and backward. Transmitting. In the fifth step, the third message is compared by comparing the second identification information extracted from the third message with the identification information on the node that is the starting point of the third message in the node list. It is determined whether or not has been received.

  In the second invention, the identification information may include a message ID assigned to each message in node units and a node ID unique to the node. In this case, it is preferable that the fourth step includes a step of adding, to the second message, a newly-numbered message ID and its own node ID indicating that it is the starting point as identification information. Moreover, it is preferable that a 5th step includes the step which specifies the node used as the starting point of the 3rd message based on node ID extracted from the 3rd message.

  According to the first or second invention, the status indicating whether or not the node is connected in the communication path in the ring is managed for each node. When the first message does not reach the destination node, the node that has detected the first message transmits the second message to notify the other nodes of the fact. Since this second message is transmitted and transferred in both directions, even if a part of the communication path in one transfer direction is disconnected, as long as the communication path in the other transfer direction is maintained, the entire message is transmitted to the entire ring. The message is surely distributed. Therefore, the reliability of the communication path in the ring can be improved. At the same time, the transfer efficiency of messages in the ring is improved, and the overall communication time required for message sharing in the ring can be reduced. As a result, management information necessary for maintaining the ring can be quickly and reliably shared throughout the ring.

  FIG. 1 is a configuration diagram of a bidirectional ring network according to the present embodiment. In this network, n (n ≧ 3) nodes 1 (1) to 1 (n) are configured in a ring shape. Each node corresponds to a single network communication device, and in this embodiment, a cellular phone having a P2P (Peer to Peer) communication function based on TCP / IP is used. A characteristic of this network is that a node that detects the disconnection of a node becomes a message source (hereinafter referred to as “origin”), and that message is bidirectional in the ring, that is, forward and reverse. The point is to multicast the message in the ring by relaying on. In the following description, the “forward direction” is referred to as a clockwise direction, and the “reverse direction” is referred to as a counterclockwise direction. However, there is no essential meaning in the reverse direction, and this is merely determined for convenience of description. (Thus, the reverse order may be defined in the opposite sense).

  In this ring network, an arbitrary node 1 (1) to 1 (n) is the starting point, and a notification message (indicating that a certain node has been updated to a disconnected state) is sent as a starting point. The notification message transfer procedure starting from the node 1 (1) will be schematically described with reference to FIGS. First, as shown in FIG. 2, the node 1 (1), which is the source of some message, confirms whether or not the message has arrived at the destination node 1 (2) when transmitting this message. . Then, the node 1 (2) determined that the message has not arrived is regarded as a disconnected state. The node 1 (1) generates a notification message to that effect in order to notify each node in the ring (except for the node 1 (2)) that the node 1 (2) has been disconnected.

  Next, as shown in FIG. 3, the origin node 1 (1) skips the node 1 (2) that has not been reached, and the nearest node 1 after the node 1 (2) in the forwarding path in the ring. In response to (3), a notification message is transmitted in the forward direction. At the same time, the node 1 (1) transmits the same notification message to the adjacent node 1 (n) in the reverse direction at the same time. As a result, the transfer of the notification message transmitted in the forward direction and the transfer of the notification message transferred in the reverse direction coexist in the ring.

  The notification messages transmitted in the forward direction are sequentially transferred through a communication path as shown in FIG. First, the node 1 (3) that has received the notification message from the origin node 1 (1) determines whether or not to transfer it to the next node, that is, the downstream node in message transfer (transfer determination). The forward transfer of the notification message is performed on the condition that the own node has not yet received the same message from the reverse direction. When the notification message has been received, the forward transfer ends at the own node. Regarding node 1 (3), at the time when the node 1 (3) has received the forward notification message, it has not yet received the reverse notification message. Therefore, node 1 (3) knows that node 1 (2) has been disconnected by the received notification message, and forwards the notification message to node 1 (4) adjacent in the forward direction. To do. Thereafter, as long as the determination results of the nodes 1 (4) to 1 (6) are not received in the reverse direction, the notification message is sent to the node 1 (4) → 1 (5) → 1 (6) → 1 (7). Are sequentially transferred through the communication path. Simultaneously with the transition of such forward transfer, the reverse transfer also shifts through a different communication path. Therefore, before any node receives the notification message from the forward direction, a state in which the same message is received from the reverse direction (transfer determination = reception received in the reverse direction) occurs. Since the node 1 (7) corresponding to the received node does not transmit the received notification message, the series of forward transfers terminates at the node 1 (7).

  On the other hand, the notification messages transmitted in the reverse direction are sequentially transferred through a communication path as shown in FIG. First, the node 1 (n) that has received the notification message from the origin node 1 (1) determines whether or not to transfer it to the next node. The reverse transfer of the notification message is performed on the condition that the own node has not yet received the same message from the forward direction. If the notification message has been received, the reverse transfer is completed at the own node. Regarding node 1 (n), when it has received a notification message from the reverse direction, it has not yet received the notification message from the forward direction. Therefore, the node 1 (n) knows that the node 1 (2) has been disconnected by the received notification message, and then sends a notification message to the adjacent node 1 (n-1) in the reverse direction. Forward. Thereafter, as long as the determination results of the nodes 1 (n-1) to 1 (n-4) are not received in the forward direction, the notification message is sent from the node 1 (n-1) → 1 (n-2) → 1. The data is sequentially transferred through the communication path of (n-3) → 1 (n-4). This reverse transfer terminates when it reaches the node 1 (n-4) that has received the same notification message in the forward transfer.

  Since it is guaranteed that the notification message is distributed in the reverse direction after node 1 (6), which is the end of the forward transfer, there is no need to continue the forward transfer. Further, since it is guaranteed that the notification message is distributed in the forward direction after node 1 (n-4), which is the end of the reverse transfer, there is no need to continue the reverse transfer. That is, in the ring excluding the destination node 1 (2) in the disconnected state, the notification message is distributed throughout the ring when both forward transfer and reverse transfer reach the end. In such bi-directional transfer, forward transfer and reverse transfer are performed in parallel, so that the number of notification message transfers can be halved compared to unidirectional transfer.

  FIG. 6 is a configuration diagram of a single network communication device 1 corresponding to each node, and all the nodes 1 (1) to 1 (n) have a common configuration. The network communication device 1 includes four transmission / reception units 2 to 5 corresponding to a communication unit, a communication control unit 6, a node management unit 7, a starting point control unit 8, an arrival confirmation unit 9, and a transfer determination unit 10. And the transfer control unit 11. The network communication device 1 is preinstalled with an application 12 that operates in conjunction with these. As the application 12, for example, a battle network game or the like that needs to share a message in real time can be used. The network communication device 1 has a function of performing normal message communication, a function as a starting point of a notification message transmitted when a node is disconnected, and a function of transferring a notification message (including a transfer termination function). .

  The forward receiving unit 2 receives a message from a node that is a message transfer source when a forward message is transferred in the ring. The forward transmission unit 3 transmits a message to a node that is a message transfer destination during forward transfer. On the other hand, the reverse direction receiving unit 4 receives a message from a node that is a message transfer source when a reverse message is transferred in the ring. The reverse transmission unit 5 transmits a message to a node that is a transfer destination of the message at the time of reverse transfer. A message to be transmitted to another node is transmitted after being packetized into a TCP / IP-compliant format by the transmission units 3 and 5, but the message and the packet are not necessarily in a one-to-one relationship. A plurality of messages may be made into one packet. Similarly, the process of extracting a message from a packet received from another node is performed by the receiving units 2 and 4, but a plurality of messages may be extracted from one packet. In addition, these transmission / reception parts 2-5 are implement | achieved as one function of the communication unit (communication part) which transmits / receives a message between other nodes in hardware.

  The communication control unit 6 transmits / receives a normal message (in contrast to the notification message, mainly due to a request from the application 12) to / from other nodes. The communication form of the normal message may be any form, for example, multicast by unidirectional ring communication or bidirectional ring communication, or unicast (typically, communication between specific nodes in the ring) Relay transfer in the ring is used). The communication control unit 6 operates when a normal message is transmitted from the application 12 to another node in response to a call origination request, or when a normal message received from another node is transferred.

  The node management unit 7 manages the message received by its own node by forwarding within the ring for each node that is the starting point of the message. Specifically, unique identification information for identifying a message transferred in the ring is stored and managed for each starting point of the message. The node management unit 7 plays an important role in determining the transfer described above, that is, whether or not its own node has already received the same message as the received message. Have. In this embodiment, in addition to the received message, a message transmitted from its own node is also managed by the node management unit 7, but this is indispensable when considering the main features of the present invention. is not.

  FIG. 7 is an explanatory diagram of a node list as an example. In this node list, a list order, an IP address, a status, and a sequence number are described in association with the node ID. Here, the “node ID” is node-specific information assigned to each node in order to identify each node in the ring. The “IP address” describes an address used when the network communication device 1 corresponding to each node communicates on the TCP / IP network. When the IP address is changed due to movement of the network communication device 1 or the like, the description content of the “IP address” is also changed accordingly. The “list order” defines a communication path between nodes in the ring, and indicates that communication is performed between the adjacent list orders. For example, the node ID “100001” with the list order “1” communicates with the node ID “100002” with the list order “2”. Note that the node with the lowest list order is handled as communicating with the highest node. A ring is defined by a set of communication paths between such nodes. “Status” indicates the connection status of the node in the corresponding column. The currently connected node is set to “connected”, and the node that has been disconnected for some reason is set to “disconnected”. . In “sequence number”, a sequence number related to the latest message transmitted from the node in the corresponding column is described. The sequence number is one of message IDs that are individually assigned to individual messages in units of nodes, and is assigned in ascending order so that the numbers are continuous. Since this numbering is performed independently for each node, duplicate numbering of the same number is allowed in a plurality of nodes. In this embodiment, a set of a sequence number as a message ID and a start node ID that specifies a start node ID is used as the identification information, whereby the start node identification and message identification are performed. It is possible.

  The node list is distributed and shared to all nodes by a matching server (not shown) when all nodes participating in the ring network are determined. In the initial state, the “status” of all the nodes is set to “connected”, and this is changed to “disconnected” when the disconnection of the node occurs after the fact. In the initial state, all “sequence numbers” are set to “0” (indicating that no message is transmitted from the origin), and are incremented every time the message is transmitted from the origin. In addition, even when a message is received by forwarding, it is updated as necessary with the sequence number extracted from the message as necessary. The same applies to the “IP address”, and is changed as needed when the IP address of the node is changed due to movement of the area of the node or the like. The node that detects the change in the description contents notifies update information indicating the change contents of the node list in a message format to all other nodes. Thereby, information (status etc.) required in order to maintain a link is shared within a ring. Note that the matching server is involved only in the initial matching, and is not involved in intra-ring communication after distribution of the node list.

  The arrival confirmation unit 8 operates when a normal message is transmitted by the communication control unit 6, or when a notification message is transmitted by the origin control unit 9 or the transfer control unit 11, and the message to be transmitted is transmitted. It is determined whether or not the destination has been reached (arrival confirmation). If the message to be sent does not arrive at this destination, update the status of the destination ("Connect" → "Disconnect"), skip the destination, and transfer routes in the ring An attempt is made to transmit the message to be transmitted to the nodes after the transmission destination. At the same time, a notification message to the effect that the status relating to the disconnected transmission destination has been updated is transferred in both directions within the ring from its own node.

  The starting point control unit 9 operates when transmitting a starting point, that is, when a notification message indicating that the status of a certain node has been changed to “disconnected” starts from its own node. In this case, the identification information (management contents) regarding the node itself in the node management unit 7 is updated to the identification information numbered to be added to the notification message. Then, the forward direction transmission unit 3 and the reverse direction transmission unit 5 are controlled by the origin control unit 9, and a notification message with identification information is transmitted in both directions.

  The transfer determination unit 10 and the transfer control unit 11 operate at the time of transfer reception, that is, when the own node receives a notification message. In this case, the transfer determination unit 10 compares the identification information extracted from the notification message from a certain starting point with the management content of the identification information related to this starting point in the node management unit 7, and whether or not the notification message has been received. Is determined. Then, according to the determination result, one of the forward transmission unit 3 and the backward transmission unit 5 corresponding to the transfer direction is controlled by the transfer control unit 11. Specifically, when it is determined that the notification message has not been received (in the case of a new notification message that has not yet been received), the node management unit 7 is accessed, and the identification information regarding the origin from which this notification message was transmitted is Updated to the identification information of the received notification message. As a result, the notification message is handled as having been received in its own node thereafter. At the same time, the node management unit 7 updates the status regarding the node notified by the notification message to the disconnected state. Then, the transfer of the notification message in this transfer direction is permitted. On the other hand, when it is determined that the same notification message has been received (when the same message is received repeatedly), transfer of the notification message in this transfer direction is prohibited. In this case, the node management unit 7 does not update the identification information and the status.

  FIG. 8 is a flowchart of communication control executed in the network communication device 1. First, in Step 1, the arrival confirmation unit 8 determines whether or not a message (including both a normal message and a notification message) transmitted (or forwarded) from its own node has reached the destination (or forwarding destination). Confirm arrival. Various methods of arrival confirmation are known, and the contents thereof are not described in detail, but typically the following methods are well known. In the starting point control, arrival confirmation is separately performed for each of the forward transmission destination and the reverse transmission destination.

(1) Method for Acquiring Transmission Information of TCP Protocol Counting time triggered by notification from control units 6, 9, and 11 (notification from transmission units 3 and 5), and transmission of TCP protocol within a predetermined time If the node cannot acquire the information, it is considered that the packet has not reached the destination.
(2) Method for Acquiring Transmission Information of Data Link Layer Protocol such as PPP (Point to Point Protocol) When the node cannot acquire transmission information of data link layer protocol such as PPP within the predetermined time, It is assumed that the packet did not reach the destination.
(3) Method of using survival confirmation PING (Packet Internet Groper) periodically transmitted from the destination When the survival confirmation is not completed within the predetermined time, it is considered that the packet has not reached the destination. .

  If the transmission message has not yet arrived, the starting point control of step 2 is executed via the branch of step 1 in order to share management information (node list update information) necessary for maintaining the ring within the ring. . On the other hand, when the notification message is received, the transfer control with exception of step 4 is executed through the branch of steps 1 and 3. Here, “exception” in transfer control with exception means that there is a case where message transfer is not performed (when transfer is terminated) depending on the result of transfer determination.

(Starting point control)
FIG. 9 is a detailed flowchart of step 2 in FIG. 8, that is, the starting point control executed when the starting point of the notification message is transmitted. First, in step 10, the origin control unit 9 accesses the node list and changes the status regarding the destination node (unreachable node) to which the message has not arrived from “connected” to “disconnected”. At the same time, the starting point control unit 9 generates a notification message including update information indicating that the status of the unreached node has been changed.

  In subsequent step 11, the starting point control unit 9 assigns a sequence number to be assigned to a new notification message to be transmitted this time in ascending order. Numbering of sequence numbers in its own node is performed regardless of the numbering status of other nodes. In subsequent step 12, the starting point control unit 9 updates the sequence number (description content) related to its own node in the node list to the sequence number assigned in step 11. For example, in FIG. 6, when the node ID “100003” performs the origin transmission, the newly assigned “1124” is overwritten on the sequence number “1123” corresponding to this node ID.

  In step 13, the starting point control unit 9 refers to the node list and specifies the message transmission destination in both directions. That is, as the forward transmission destination, a node closest to the self node on the ascending order of the list order and having a status of “connected” is specified. At the same time, as the transmission destination in the reverse direction, the node closest to the self node on the descending order of the list order and having the status of “connected” is specified. For example, when the node ID “100003” of the list rank “3” has the node list shown in FIG. 7 and the node ID “100002” is not yet reached (the status of this ID is changed to “disconnected”) The transmission destination in the forward direction is the node ID “100004” in the list order “4”, and the transmission destination in the reverse direction is the node ID “100001” in the list order “1” (list order “2”). Since node ID “100002” is in a disconnected state, it is skipped). Then, the message is converted into a predetermined format and then output to the forward transmission unit 3 and the reverse transmission unit 5. This message corresponds to a data set surrounded by a thick line in FIG. The header of the message includes the sequence number assigned in step 11, the message type (application message / ring management message distinction), the origin node ID (own node ID), and the destination node ID (all nodes as destinations). In such cases, the designation) is included. The payload corresponds to node list update information (unreachable node = “disconnect”).

  In the present embodiment, TCP / IP is assumed, but it is actually only the following two timings that the connection destination needs to be specified in TCP. At the time of transmission other than these, it is only necessary to read / write from / to the already established TCP stream (communication path), so it is not necessary to refer to the node list every time.

(1) When connecting for the first time (when building a ring)
(2) When previous or next node is disconnected

  In step 14, the forward transmission unit 3 and the reverse transmission unit 5 generate a packet that conforms to the TCP / IP shown in FIG. Specifically, a message received from the origin control unit 9 is used as a payload, and a header such as an IP header or a TCP header is added thereto. Then, the forward transmission unit 3 transmits a TCP / IP packet to the transmission destination in the forward direction specified by the origin control unit 9. At the same time, the reverse transmission unit 5 transmits the TCP / IP packet to the transmission destination in the reverse direction designated by the origin control unit 9.

  In step 15, the arrival confirmation unit 8 uses the arrival confirmation method described above to determine whether or not the packet has arrived at the transmission destination using the origin control unit 9 (notification from the transmission units 3 and 5 is also possible) as a trigger. judge. If the determination in step 15 is affirmative, that is, if the packet reaches the transmission destination, a series of starting point control for the message that is the current transmission target ends. On the other hand, if a negative determination is made in step 15, that is, if the packet does not reach the transmission destination, the process proceeds to step 16. In response to the determination result from the arrival confirmation unit 8, the starting point control unit 9 changes the status regarding the unreachable transmission destination in the node list from “connected” to “disconnected”. Then, returning to step 13, the origin control unit 9 refers to the node list, selects and specifies a new transmission destination for the unreachable direction, and then attempts to transmit a message to the new transmission destination. . By updating the status in the previous step 16, the initial transmission destination is handled as a disconnected state, so the initial transmission destination is skipped, and the node whose status is “connected” after that is selected as the new transmission destination. Is done.

(Transfer control with exception)
FIG. 11 is a detailed flowchart of step 4 in FIG. 8, that is, the transfer control with exception executed when receiving the notification message. First, in step 21, the transfer determination unit 10 determines whether or not the notification message received this time has been received. Specifically, the sequence number extracted from the received message is compared with the description content (sequence number) of the node list regarding the node ID extracted from the received message. If both match, this means that a message received this time in a certain transfer direction has already been received in the opposite transfer direction. Therefore, if the determination result in step 21 is affirmative (when the same message is received in duplicate), the process ends without transmitting the received message to avoid duplicate transfer. On the other hand, if the determination result in step 21 is negative (when an unreceived message is received for the first time), the process proceeds to step 22 to transfer the received message.

  In step 22, the transfer control unit 11 updates the node list of its own node based on the notification message received this time. Specifically, first, in the node list, the description content of the sequence number related to the origin node is updated to the sequence number extracted from the received message. For example, in FIG. 7, when the origin node ID extracted from the message received by the node ID “100003” is “100001” and the sequence number is “920”, the sequence number “919” corresponding to this node ID is set to “920”. "Is overwritten. As a result, the node ID “100003” is handled as “already received” up to “920” with respect to the message from the origin node ID “100001”. Secondly, the status regarding the node (unreachable node) notified by the notification message is changed to “disconnected”. Thereby, the update information transmitted from the origin node is reflected in this node.

  In step 23, the transfer control unit 11 refers to the node list and specifies a message transfer destination (transmission destination). In the message transfer, the transfer direction of the received message is maintained as it is. In the case of forward transfer, a node that is closest to its own node on the ascending order of the list order and has a status of “connected” is specified as a forward transfer destination. On the other hand, in the case of reverse transfer, a node closest to the self node on the descending order of the list order and having a status of “connected” is specified as the transfer destination in the reverse direction. Then, the message is converted into a predetermined format and output to one of the forward transmission unit 3 and the reverse transmission unit 5 corresponding to the transfer direction. The format of this message is the same as that generated by the origin control unit 9. In step 24, one of the forward transmission unit 3 and the reverse transmission unit 5 (one corresponding to the transfer direction) generates a TCP / IP-compliant packet and transmits the TCP / IP packet.

  In step 25, the arrival confirmation unit 8 determines whether or not the packet has reached the transfer destination using the above-described arrival confirmation method. If the determination in step 25 is affirmative, that is, if the packet has reached the transfer destination, a series of transfer controls relating to the notification message that is the current transmission target ends. On the other hand, if a negative determination is made in step 25, that is, if the packet does not reach the transfer destination, the process proceeds to step 26. In response to the determination result from the arrival confirmation unit 8, the transfer control unit 11 changes the status regarding the unreachable transfer source in the node list from “connected” to “disconnected”. And it returns to step 23 and the procedure of steps 23-25 is performed again. As a result, a new transfer destination is selected and specified for the transfer direction that has not been reached, and transmission of a message to this transfer destination is attempted. This new transfer destination is not the original transfer destination but a node whose status is “connected” after this.

  As described above, in this embodiment, the status indicating whether or not the node is connected in the communication path in the ring is managed for each node, and based on this, the message received in one of the forward direction and the reverse direction ( It is determined whether or not to transfer (indicating that a certain node is disconnected). As a result, the forward transfer and the reverse transfer related to a certain message are terminated at a node that has received the same message, and are not transferred redundantly to subsequent nodes. Accordingly, the transfer efficiency of messages in the ring is improved, and the overall communication time required for message sharing in the ring can be reduced. At the same time, the reliability of the communication path in the ring can be improved. Because the message is relayed in both directions, even if a part of the communication path in one transfer direction is disconnected, the message is surely transmitted to the entire ring as long as the communication path in the other transfer direction is maintained. It is because it goes around. As a result, according to the present invention, management information necessary for maintaining the ring can be quickly and reliably shared throughout the ring.

  In the embodiment described above, a mobile phone capable of P2P communication has been described as an example of the network communication device. In the services currently provided by mobile phone carriers, there are restrictions on the number of transmissions per unit time, which is a bottleneck for multicasting. The present invention is one of effective techniques that can easily realize multicast on a mobile phone with such restrictions. However, the application target of the present invention is not limited to a mobile phone, and can be widely applied to various mobile terminals and even general computers (usually fixedly installed). .

  In the above-described embodiment, a case has been described in which a single node management unit (node list) is used to manage the connection state of each node and manage received messages. However, it is not necessary to manage them centrally. For example, both may be managed separately by preparing a list for each application.

  In the above-described embodiment, a case has been described in which a set of a sequence number and a node ID is used as identification information for identifying a message transferred in the ring. This sequence number has the following two characteristics.

(1) Numbering for each message in node units (2) Numbering so that numbers are consecutive (sequential numbering)

  Regarding (1) above, the reason for the node unit is to eliminate the communication for inquiring the numbering status, which is necessary when the unit is a ring. When assigning numbers so that they are unique within the ring (when multiple nodes do not allow duplicate numbers), the node that is about to assign a new number inquires about the current numbering status at other nodes. Must. Communication for this inquiry becomes a big bottleneck in a mobile phone with a large restriction on communication resources. On the other hand, in the case of a node unit, each node can arbitrarily assign a number regardless of the numbering situation of other nodes, so that such communication can be eliminated. Of course, if the restriction of communication resources hardly becomes a problem, numbering may be performed in units of rings, and the present invention does not exclude such a form.

  In the case of numbering in units of nodes, it is necessary to adopt a mechanism that allows the node that received the message to identify the starting node. As this mechanism, there is a method of adding a starting node ID to a message as in the above-described embodiment, but in addition to this, the starting node can be specified by providing a fixed rule for sequence numbering itself. You may do it. For example, the remainder of the sequence number is associated with the starting node. Specifically, when the number of nodes in the ring is 4, the nodes in the list order “1” are assigned sequence numbers in ascending order of remainder 1, such as 1, 5, 9, 13,. Similarly, the node with the list rank “2” is the remainder 2, the node with the list rank “3” is the remainder 3, the node with the list rank “4” is the remainder 0, and each is assigned a sequence number. Thereby, each node can identify the starting node from the remainder of the sequence number extracted from the message. As another method, the node order may be identified by setting the list order of the node list to an appropriate number of bits of the sequence number.

  In addition, regarding (2) above, the reason for sequential numbering is to suppress the data amount of the node list. When sequential numbering is adopted as in this embodiment, the progress of messages in the node list can be managed only by the description of the latest sequence number, and all past sequence numbers are recorded in history (starting transmission history / received history). ) Is not necessary (and therefore can be overwritten). This is because messages before the latest sequence number can be handled as received. Of course, if the data amount of the node list is not taken into consideration, all sequence numbers can be managed as history, and the present invention does not exclude such a form. In the case of management in a history format, it is not necessary to use a sequence number, and it is sufficient to use a simple message ID unique to the message (this message ID is assigned in units of nodes or rings).

  Furthermore, in the above-described embodiment, the description has focused on the functional configuration and operation of the network communication device 1. However, a bidirectional ring network communication program for causing the network communication device 1 to realize such a function is also provided. It constitutes another embodiment of the invention. By installing this communication program, generally, the network communication device 1 can execute the following steps (details are omitted because they overlap with the description of the above-described embodiment). It should be noted that these steps do not indicate a chronological order, but are merely arranged by function.

(1) Preparation of node list (2) Linkage with application (3) Update of node list (4) Message transmission / reception with other nodes (5) Monitoring of node connection status (6) Update information of node list Origination transmission and transfer

Ring network configuration diagram Message unreachable illustration Explanatory diagram of message transmission by origin Illustration of message transfer in forward transfer Illustration of message transfer in reverse transfer Configuration diagram of network communication equipment Illustration of node list as an example Communication control flowchart Detailed flowchart of starting point control Explanation of TCP / IP compliant packet Detailed flowchart of transfer control with exception

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Network communication apparatus 2 Forward direction reception part 3 Forward direction transmission part 4 Reverse direction reception part 5 Reverse direction transmission part 6 Communication control part 7 Node management part 8 Arrival confirmation part 9 Origination control part 10 Transfer determination part 11 Transfer control part 12 Application

Claims (12)

A network in which a plurality of nodes are configured in a ring and used as the node in a ring network that multicasts a message in a ring by transferring a message transmitted from an arbitrary node in the ring. In communication equipment,
A communication unit for sending and receiving messages to and from other nodes;
A first node management unit that manages, for each node, a status indicating whether or not a node is connected in a communication path in the ring;
An arrival confirmation unit that determines whether or not the first message transmitted from its own node via the communication unit has reached the destination node;
If it is determined that the first message has not arrived at the destination node, the node management unit is accessed to update the status related to the node to a disconnected state, and the node is updated to a disconnected state. By transmitting a second message to the effect in the forward direction and the reverse direction in the ring excluding the node via the communication unit, the status update content is transmitted to each node in the ring. A network communication device comprising a starting point control unit for notification. A second node management unit that manages a message received by its own node by forwarding in the ring for each node that is a starting point of the message;
When the third message transmitted by a certain starting point is received by the communication unit in either the forward direction or the reverse transfer direction, the second node management unit is accessed, and the third message A transfer determination unit that determines whether or not the third message has been received with respect to the starting node;
When it is determined that the third message has not been received, the first node management unit is accessed to update the status regarding the node notified by the third message to a disconnected state, and 2, the identification information related to the node that is the starting point of the third message is updated, and the third message is transmitted in the transfer direction within the ring via the communication unit. A transmission control unit that transmits and terminates the transfer of the third message in the transfer direction at its own node without transmitting the third message when it is determined that the third message has been received. The network communication device according to claim 1, further comprising: The second node management unit manages identification information for identifying a message transferred in the ring for each node that is a starting point of the message,
The starting point control unit numbers the first identification information to be given to the second message, transmits the second message to which the first identification information is given in the forward direction and the reverse direction,
The transfer determination unit compares the second identification information extracted from the third message with the identification information on the node that is the starting point of the third message in the node management unit, thereby comparing the second identification information. 3. The network communication device according to claim 2, wherein it is determined whether or not the message 3 has been received.   4. The network communication device according to claim 3, wherein the identification information includes a message ID numbered in units of nodes for each message and a node ID unique to the node.   The origin control unit assigns, to the second message, a newly assigned message ID and an own node ID indicating that the origin is a starting point, using the identification information. 4. The network communication device described in 4.   6. The network communication device according to claim 5, wherein the transfer determination unit identifies a node that is a starting point of the third message based on a node ID extracted from the third message. . A network in which a plurality of nodes are configured in a ring and used as the node in a ring network that multicasts a message in a ring by transferring a message transmitted from an arbitrary node in the ring. In a communication program installed in a communication device,
A first step of preparing a node list for managing, for each node, a status indicating whether or not a node is connected in a communication path in the ring;
A second step of determining whether or not the first message transmitted from its own node has reached the destination node;
If it is determined that the first message has not arrived at the destination node, a third step of updating the status of the node in the node list to a disconnected state;
By transmitting a second message indicating that the node has been updated to the disconnected state in the forward direction and the reverse direction in the ring, the status update content is transmitted to each node in the ring excluding the node. A communication program for a ring network, comprising: a fourth step of notifying. The node list manages the message received by its own node by forwarding in the ring for each node that is the origin of the message,
When a third message transmitted by a certain origin is received in either the forward or reverse transfer direction, the third message is related to the node that is the origin of the third message in the node list. A fifth step of determining whether or not it has been received;
When it is determined that the third message has not been received, the status related to the node notified by the third message in the node list is updated to a disconnected state, and the third message in the node list is updated. A sixth step of transmitting the third message in the transfer direction in the ring after updating the identification information relating to the originating node;
If it is determined that the third message has been received, a seventh step of terminating the transfer of the third message in the transfer direction at its own node without transmitting the third message; The ring network communication program according to claim 7, further comprising: The node list manages identification information for identifying a message transferred in the ring for each node that is a starting point of the message,
The fourth step includes
Numbering first identification information to be given to the second message;
Transmitting the second message with the first identification information forward and backward, and
The fifth step includes
By comparing the second identification information extracted from the third message with the identification information regarding the node that is the starting point of the third message in the node list, the third message has been received. 9. The ring network communication program according to claim 8, which is a step of determining whether or not there is a ring network communication program.   10. The ring network communication program according to claim 9, wherein the identification information includes a message ID assigned to each message in node units and a node-specific node ID.   The fourth step includes a step of assigning, as the identification information, a newly assigned message ID and a self node ID indicating that it is a starting point to the second message. The ring network communication program according to claim 10.   The method according to claim 11, wherein the fifth step includes a step of identifying a node that is a starting point of the third message based on a node ID extracted from the third message. A ring network communication program.

Images