JP2005086525A - Communication system in ad hoc network - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a communication system for appropriately controlling TCP communication in accordance with link disconnection or new link connection in an ad hoc network for using a table drive type routing protocol to perform routing. <P>SOLUTION: When a terminal 1 starts TCP communication, a transmission destination address and a transmission source port number are registered in a socket information management table, when a terminal 3 detects link disconnection, Metric of path information of a routing table is updated to BIG and an Update packet including the updated path information is broadcasted, and when a terminal 2 and the terminal 1 receive the Update packet, the Metric is updated to BIG. When the Metric becomes the BIG, the terminal 1 acquires the transmission destination address of the path information, refers to the socket information management table and freezes TCP communication using a transmission source port number corresponding to the acquired transmission destination address. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a communication method in an ad hoc network, and more particularly to a communication method in an ad hoc network that performs routing using a table-driven routing protocol.

  An ad hoc network is a network that is autonomously and instantaneously constructed by a plurality of wireless terminals communicating with each other. In an ad hoc network, when two wireless terminals that communicate with each other do not exist in each other's communication area, a wireless terminal located between the two wireless terminals functions as a router and relays data packets, thereby forming a wide-range multi-hop network. can do.

  In an ad hoc network, link disconnection frequently occurs due to movement of a wireless terminal to be relayed. Therefore, it is necessary to quickly reconstruct the route, and various routing protocols have been proposed.

  The routing protocols for ad hoc networks proposed to date can be divided into two groups.

  The first group is an on-demand routing protocol. In this group of routing protocols, every time a communication request occurs (on demand), a route to the destination is searched. A typical protocol is the Dynamic Source Routing (DSR) protocol.

  The second group is a table driven routing protocol. In this group of routing protocols, a route to each destination is created in advance as a table entry. As a typical protocol, there is a Destination-Sequenced-Distance-Vector (DSDV) protocol (see, for example, Non-Patent Document 1).

  By the way, many of the applications operated on the wired network use TCP as a transport layer protocol. Therefore, in order to operate an existing application even on an ad hoc network, it is necessary to use TCP as a transport layer protocol.

  In the TCP communication, even if the link is disconnected somewhere between the terminals of the TCP communication that established the connection, the TCP communication is continued, a useless packet is transmitted, and the throughput of the TCP communication is reduced.

As a method for dealing with this problem, in Non-Patent Document 2, for DSR which is an on-demand type routing protocol, TCP communication is frozen when a link is disconnected, and TCP communication is performed when a link is newly connected. A TCP communication control method for releasing the freeze is described.
CEPerkins and P. Bhagwat, "Destination sequenced distance-vector rooting (DSDV) for mobile computers", Proc. ACM / IEEE SIGCOM94, pp234-244, Aug. 1994 G. Holland and NHVaidya, "Analysis of TCP performance over middle ad hoc networks", Proc. ACM / IEEE MOBICOM'99, pp219-230, Aug.1999

  However, the method described in Non-Patent Document 2 described above is a method specific to an on-demand type routing protocol. Such a method cannot be used for a table-driven routing table.

  In the on-demand routing protocol, the route search is performed in response to a packet transfer request, so that the start of packet transfer is delayed, whereas the table-driven routing protocol does not have such a problem. In the table driven routing protocol, route update information is transmitted to the entire network in response to a change in the topology of the network.

  Thus, in a table-driven routing protocol that has the advantage of being able to quickly start forwarding packets and respond quickly to dynamic changes in topology, when a link is broken or a new link is connected Sometimes it is important to control TCP communication.

  SUMMARY OF THE INVENTION Therefore, an object of the present invention is to provide a communication system in an ad hoc network that appropriately controls TCP communication according to link disconnection or new link connection in an ad hoc network that performs routing using a table-driven routing protocol. Is to provide.

  In order to solve the above-described problems, the present invention is a communication method in an ad hoc network that performs routing using a table-driven routing protocol, in which a TCP communication source terminal establishes a TCP communication connection when a TCP communication connection is established. Registering the socket information including the address of the communication destination terminal, the source port number, and the destination port number in the socket information management table, and when each terminal detects link disconnection, The step of updating the corresponding route information metric in the routing table to a value or symbol indicating link disconnection, broadcasting a packet including the updated route information, and each terminal receiving the packet including the updated route information When the corresponding routing table of its own terminal The information metric is updated to a value or symbol indicating link disconnection, the packet including the updated route information is broadcasted, and the transmission source terminal of the TCP communication has the metric of the route information in the routing table of its own terminal When a link disconnection value or symbol is obtained, the address of the destination terminal of the route information is acquired from the routing table, and the socket information including the acquired address of the destination terminal is identified by referring to the socket information management table And freezing TCP communication using the source port number included in the specified socket information.

  Preferably, the socket information further includes a status indicating whether the state of the TCP communication is frozen or an action indicating a normal operation state. In the step of registering the socket information, the source terminal of the TCP communication In the step of registering the socket information further including the status indicating the action in addition to the address, the transmission source port number, and the transmission destination port number in the socket information management table, and freezing the TCP communication, the transmission communication terminal of the TCP communication Only when the status of the socket information is an action, the TCP communication using the source port number of the socket information is frozen, and the source terminal of the TCP communication updates the status of the socket information to freeze after the TCP communication is frozen. The method further includes a step.

  More preferably, in the communication method in the ad hoc network, when each terminal discovers a new route, the routing table is updated with the route information of the new route with the metric as the number of hops, and the updated route information is broadcast. The transmission step, and when the metric of the route information in the routing table of its own terminal changes from a value or symbol indicating link disconnection to the number of hops, the destination terminal of the route information from the routing table The socket information management table is referenced to identify the socket information including the address of the acquired destination terminal. The source of the identified socket information is only when the status of the identified socket information is frozen. Freeze state of TCP communication using port number A step of dividing, the source terminal of the TCP communication, after release of the freeze state of the TCP communication, and updating the status of the socket information to the action.

  More preferably, the table driven routing protocol is the DSDV protocol.

  Preferably, when the TCP communication source terminal disconnects the TCP communication connection, the socket information including the address of the TCP communication destination terminal, the source port number, and the destination port number is deleted from the socket information management table. The method further includes a step.

  The present invention is also a communication method in an ad hoc network that performs routing using a table-driven routing protocol, and an IP module that executes IP of a TCP communication source terminal executes TCP processing. Registering socket information including a TCP communication destination terminal address, a source port number, and a destination port number in the Internet layer socket information management table when establishing a TCP communication connection by When the daemon that executes the table-driven routing protocol receives a notification from the MAC module that executes the MAC protocol that the link has been disconnected together with the packet that has not been transferred, the routing table of its own terminal Corresponding A step of updating the routing information metric to a value or symbol indicating link disconnection, broadcasting a packet including the updated routing information by a UDP module executing the UDP protocol, and the daemon of each terminal updating the routing information Updating the corresponding route information metric in the routing table of its own terminal to a value or symbol indicating link disconnection when the packet including the route information is received, and broadcasting the packet including the updated route information by the UDP module; When the TCP communication source terminal daemon monitors the routing information metric in the routing table of its own terminal and the routing information metric becomes a value or symbol indicating link disconnection, the routing information from the routing table Of the destination device TCP address using the source port number included in the specified socket information is acquired for the TCP module by specifying the socket information including the address of the acquired destination terminal by acquiring the address and referring to the socket information management table Freezing the communication.

  Preferably, the step of registering the socket information includes a step of obtaining a TCP header from the IP packet when the IP module of the TCP communication source terminal transmits the IP packet, and an IP module of the TCP communication source terminal. However, when SYN is set in the Code Bit in the TCP header, the step of acquiring the IP header from the IP packet, and the IP module of the source terminal of the TCP communication are the addresses of the source terminal in the IP header. Is the address of the own terminal, the source port number and the destination port number are acquired from the acquired TCP header, the address of the destination terminal is acquired from the acquired IP header, and the address and transmission of the acquired destination terminal are transmitted. Storing the source port number and the destination port number in memory; The step of setting a flag indicating that the IP module of the source terminal of CP communication is in the process of establishing a connection to a predetermined value, and the flag of the IP module of the source terminal of TCP communication are set to a predetermined value When the IP packet is received, the step of obtaining the TCP header from the IP packet and the IP module of the TCP communication transmission source terminal are set to SYN or ACK in the Code Bit in the TCP header. Sometimes, further, the step of obtaining the IP header from the IP packet and the IP module of the TCP communication source terminal are held in the memory when the address of the destination terminal in the IP header is the address of the own terminal Including the destination terminal address, source port number and destination port number. The socket information including an action indicating the work status as a status comprising the step of registering the socket information management table.

  More preferably, the communication method in the ad hoc network further includes a step of acquiring a TCP header from the IP packet when the IP module of the TCP communication source terminal transmits the IP packet, and the TCP communication source terminal When FIN is set in the Code Bit in the TCP header, the IP module further acquires the IP header from the IP packet, and the IP module of the TCP communication source terminal is connected to the source terminal in the IP header. A source port number and a destination port number are acquired from the acquired TCP header and the address of the destination terminal is acquired from the acquired IP header, and a TCP communication source The IP module of the terminal acquires the address of the acquired destination terminal. , Source port number, and the socket information including destination port number and a step of deleting from the socket information management table.

  According to the present invention, in an ad hoc network that performs routing using a table-driven routing protocol, TCP communication can be appropriately controlled according to link disconnection or new link connection.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  FIG. 1 shows the configuration of an ad hoc network. As shown in the figure, a plurality of wireless terminals 1 to 7 autonomously form a network. With this ad hoc network, for example, the wireless terminal 3 is located away from the wireless terminal 4 and cannot directly receive the packet. Therefore, the packet transmitted from the wireless terminal 4 is transmitted to the wireless terminal 2, the wireless terminal 9, and It is relayed by the wireless terminal 6 and delivered to the wireless terminal 3.

FIG. 2 shows the configuration of the wireless terminal 1. The configurations of the other wireless terminals 2 to 9 are the same.
With reference to FIG. 2, the wireless terminal 1 includes an input unit 11, a display unit 12, an antenna 25, an e-mail application 13, and a communication control unit 14.

  The e-mail application 13 is a so-called mailer, which creates data based on input of a message, a destination, and the like from the input unit 11 and sends the data to the communication control unit 14. Further, the e-mail application 13 causes the display unit 12 to display a message based on the data received from the communication control unit 14.

  The antenna 25 receives radio signals from other radio terminals and transmits radio signals to other radio terminals.

  The communication control unit 14 includes a plurality of modules that perform communication control according to an ARPA (Advanced Research Projects Agency) Internet hierarchical structure.

  The radio interface module 24 in the physical layer performs modulation / demodulation of transmission signals or reception signals, frequency conversion, and the like according to a predetermined rule.

  The LLC module 22 in the upper layer of the network layer is an execution processing unit of the LLC protocol, and connects and releases a link with an adjacent wireless terminal.

  The MAC module 23 in the lower layer of the network layer is a MAC protocol execution processing unit, and performs retransmission control and the like. The MAC module 23 detects that the link has been disconnected because the number of packet retransmissions exceeds a predetermined value, and notifies the DSDV daemon 16 of the link disconnection and the IP packet that has not been retransmitted.

  The IP module 19 in the Internet layer is an execution processing unit for IP (Internet Protocol). The IP module 19 generates an IP packet. The IP packet includes an IP header and an IP data portion for storing a packet of a higher protocol.

  FIG. 3 shows the configuration of the IP header. The destination address and the source address in the IP header shown in the figure are used for creating socket information in the socket information management table 20.

  The TCP module 17 in the transport layer is a TCP (Transmission Control Protocol) execution processing unit. The TCP module 17 generates a TCP packet. The TCP packet is composed of a TCP header and a TCP data part for storing data of an upper protocol. This TCP packet is transferred to the IP module 19 and stored in the IP data portion in the IP packet.

  FIG. 4 shows the structure of the TCP header. The source port number in the TCP header shown in the figure is a number that identifies the application that has output this TCP packet when a plurality of applications are operating on the source wireless terminal. The transmission destination port number is a number that identifies an application that delivers this TCP packet when a plurality of applications are operating on the wireless terminal of the transmission destination. These source port number and destination port number are used for creating socket information in the socket information management table 20.

  TCP communication is an end-to-end connection-oriented communication protocol. The TCP module 17 of a terminal requesting connection connection of TCP communication (hereinafter referred to as a TCP communication connection request terminal) indicates a connection request for connection in which SYN (Synchronize Flag) is set in the Code Bit in the TCP header when the connection is established. The first packet is transmitted to the TCP module 17 of a terminal that accepts a connection connection for TCP communication (hereinafter, a TCP communication connection acceptance terminal). In response, the TCP module 17 of the TCP communication connection accepting terminal) receives the connection request acceptance and connection completion of the connection in which SYN (Synchronize Flag) and ACK (confirmation response) are set in the Code Bit in the TCP header. The packet is transmitted to the TCP module 17 of the TCP communication connection requesting terminal. Further, in response to this, the TCP module 17 of the TCP communication connection requesting terminal sends a third packet indicating the connection completion of the connection in which the Code Bit in the TCP header is set to ACK (acknowledgment response) to the TCP of the TCP communication connection receiving terminal. Transmit to module 17.

  When the TCP communication connection requesting terminal is a TCP communication source terminal (that is, a terminal that transmits data and receives an ACK for the data), the IP module 19 detects the transmission of the first packet by the TCP module 17. When the IP module 19 detects the reception of the second packet by the TCP module 17, socket information is created in the TCP communication connection request terminal.

  When the TCP communication connection receiving terminal is a TCP communication source terminal, the IP module 19 detects the transmission of the second packet by the TCP module 17 and the IP module 19 detects the reception of the third packet by the TCP module 17. As a result, socket information is created in the TCP communication connection receiving terminal.

  The connection disconnection request can be made from either the TCP communication requesting terminal or the TCP communication receiving terminal. The TCP module 17 of a terminal requesting disconnection of TCP communication (hereinafter referred to as a TCP communication disconnection request terminal) indicates a disconnection request for a connection in which the Code Bit in the TCP header is set to FIN (Finish Flag) when the connection is disconnected. The first packet is transmitted to a terminal that accepts a disconnection of TCP communication (hereinafter, a TCP communication disconnection acceptance terminal). In response to this, the TCP module 17 of the TCP communication disconnection receiving terminal sets the second packet indicating acceptance of the disconnection request for the connection in which the Code Bit in the TCP header is set to ACK (acknowledgment response), and the Code Bit in the TCP header. A third packet indicating completion of disconnection of the connection set in FIN (Finish Flag) is transmitted to the TCP module 17 of the TCP communication disconnection request terminal. In response to this, the TCP module 17 of the TCP communication disconnection requesting terminal sends a fourth packet indicating the completion of disconnection of the connection in which the Code Bit in the TCP header is set to ACK (acknowledgment response) to the TCP communication disconnection receiving terminal TCP Transmit to module 17.

  When the TCP communication disconnection requesting terminal is a TCP communication transmission source terminal, the transmission of the first packet by the TCP module 17 is detected by the IP module 19, whereby the socket information of the TCP communication disconnection requesting terminal is deleted.

  When the TCP communication disconnection receiving terminal is a TCP communication source terminal, the socket information of the TCP communication disconnection receiving terminal is deleted when the IP module 19 detects the transmission of the third packet by the TCP module 17.

  When the TCP module 17 receives an instruction to freeze the TCP communication together with the transmission source port number from the DSDV daemon 16, the TCP module 17 executes a TCP communication freeze process to be described later.

  When the TCP module 17 receives an instruction to release the freeze of TCP communication from the DSDV daemon 16 together with the transmission source port number, the TCP module freeze release process described later is executed.

  The UDP module 18 in the transport layer is a UDP (User Datagram Protocol) execution processing unit. The UDP module 18 broadcasts an Update packet created by the DSDV daemon 16, receives an Update packet broadcast from another wireless terminal, and passes it to the DSDV daemon 16.

  The process / application layer SMTP module 15 is an execution processing unit of SMTP (Simple Mail Transfer Protocol). The SMTP module 15 secures a full-duplex communication channel and exchanges messages based on the data received from the e-mail application 13.

  The process / application layer DSDV daemon 16 monitors the execution status of other communication control modules and processes requests from other communication control modules. That is, the DSDV daemon 16 is an execution processing unit for the DSDV protocol that is a table-driven routing protocol, and performs processing unique to the present embodiment.

  Hereinafter, the following processes (1) to (6), which are characteristic processes of the embodiment of the present invention, will be described. (1) Ad hoc network routing, (2) Link disconnection detection and notification, (3) Registration of socket information, (4) TCP communication freeze transition control, (5) TCP communication freeze release control, (6) Socket Delete information. Among these, (1) and (2) are processes performed in the conventional DSDV daemon, and (3) and (6) are processes unique to the IP module 19 of the present embodiment. (4) and (5) are specific processes of the DSDV daemon 16 of this embodiment.

  Hereinafter, the processing contents of (1) to (6) will be described.

(1) Routing of ad hoc network The DSDV daemon 16 periodically exchanges route information, calculates the optimum route based on the obtained route information, and dynamically creates the routing table 21 in the Internet layer. .

  FIG. 5 shows an example of the routing table 21. Referring to the figure, the route information which is each entry of this routing table 21 includes a transmission destination address, a NextHop address, a Metric, and a SeqNum.

  The transmission destination address indicates the IP address of the transmission destination terminal. The NextHop address indicates the IP address of the next hop terminal. When Metric is a predetermined value BIG, it indicates that the route (link) based on this route information is disconnected. When Metric is other than the predetermined value BIG, it indicates the number of hops to the destination terminal. Here, the predetermined value BIG is a sufficiently large value that cannot be a numerical value for representing the number of hops to the destination terminal. SeqNum (Sequence Number) indicates the order in which the route information is generated.

  In the example of the routing table shown in FIG. 5, the first route is a route in which the transmission source terminal is the wireless terminal 1 and the transmission destination terminal is the wireless terminal 3, and the packet transmitted by the wireless terminal 1 is relayed first. This indicates that the terminal is the wireless terminal 2 and the packet transmitted by the wireless terminal 1 reaches the wireless terminal 3 via three relay terminals (from Metric 4).

  The second route is a route in which the transmission source terminal is the wireless terminal 1 and the transmission destination terminal is the wireless terminal 7, and the terminal that relays the packet transmitted by the wireless terminal 1 first is the wireless terminal 5. The packet transmitted by the wireless terminal 1 is transmitted to the wireless terminal 7 via two relay terminals (from Metric 3).

  The third route is a route in which the transmission source terminal is the wireless terminal 1 and the transmission destination terminal is the wireless terminal 4, and the terminal that first relays the packet transmitted by the wireless terminal 1 is the wireless terminal 2, The packet transmitted by the wireless terminal 1 (from Metric 2) indicates that it reaches the wireless terminal 4 via one relay terminal (that is, the wireless terminal 2).

  The DSDV daemon 16 performs a periodic route search process as follows. First, the DSDV daemon 16 increments SeqNum of its route information. As a result, this route information indicates that it is newer than the route information in the routing table 21 possessed by other wireless terminals. The DSDV daemon 16 registers all the route information in the latest routing table 21 in the Update packet, and broadcasts the Update packet by the UDP module 18.

  When the DSDV daemon 16 of another wireless terminal receives this Update packet, it updates the route information in the routing table 21 of its own wireless terminal based on the route information in this Update packet, and if necessary, The UDP module 18 broadcasts an Update packet in which all the route information in the routing table 21 of the wireless terminal is registered.

  The DSDV daemon 16 performs IP packet transfer processing as follows. The DSDV daemon 16 refers to the routing table 20 and identifies the route information of the same destination address as the destination address of the generated IP packet or the IP packet received from another wireless terminal. The IP packet is transmitted to the wireless terminal specified by the NextHop address of the route information.

(2) Detection and notification of link disconnection When the DSDV daemon 16 receives a notification from the MAC module 23 that the link has been disconnected or from another wireless terminal that has received a notification that the link has been disconnected, Regardless of the periodic search process described above, the routing table 21 of the wireless terminal itself is updated, and the UDP packet is immediately broadcast by the UDP module 18 to notify other wireless terminals that the link has been disconnected. Notice.

  With reference to the flowchart shown in FIG. 6, the procedure of link disconnection detection and notification processing by the DSDV daemon 16 will be described.

  The MAC module 23 of any wireless terminal detects that the link has been disconnected by detecting that the number of retransmissions of the packet has exceeded a predetermined number by carrier sense, and the link has been disconnected. The DSDV daemon 16 is notified of the IP packet that has not been transferred (S101).

  The DSDV daemon 16 checks whether or not the destination address of the IP header of the packet that has not been transferred exists in the routing table 21, and if it exists, the destination address in the routing table 21 is included. The route information is updated to BIG (S102, S103).

  The DSDV daemon 16 creates an Update packet including the changed route information, and sends the Update packet to the UDP module 18 for broadcast transmission (S104).

  The DSDV daemon 16 of another wireless terminal receives the Update packet received by the UDP module 18 (S105).

  If the destination address of the route information included in the Update packet exists in the routing table 21 and the metric is not BIG, the DSDV daemon 16 indicates that the destination information of the route information including the destination address in the routing table 21 is included. The Metric is updated to BIG (S106, S107).

  The DSDV daemon 16 creates an Update packet including the changed route information, and sends the Update packet to the UDP module 18 for broadcast transmission (S108).

  The processes of S105 to S108 described above are performed until it is determined in S106 that the destination address of the route information included in the Update packet does not exist in the routing table 21, or even if it exists, the Metric is determined to be BIG. ,Repeated.

(3) Registration of Socket Information The IP module 19 of the TCP communication connection requesting terminal and the TCP communication connection receiving terminal transmits a terminal in TCP communication from the application 13 and receives a ACK for the data (terminal on the side ( (Hereinafter referred to as a TCP communication source terminal), the TCP communication is established based on the TCP header shown in FIG. 2 and the IP header shown in FIG. Is registered in the socket information management table 20 of the Internet layer.

  FIG. 7 shows an example of the socket information management table 20 when the TCP communication source terminal is the wireless terminal 1. Referring to the figure, socket information which is each entry in socket information management table 20 includes a transmission destination address, a transmission source port number, a transmission destination port number, and a status.

  Here, the transmission destination address indicates the IP address of the transmission destination terminal. The transmission source port number is a number that identifies an application that has output a TCP packet when a plurality of applications are operating on the transmission source wireless terminal (that is, the own terminal). The transmission destination port number is a number that identifies an application that delivers a TCP packet when a plurality of applications are operating on the wireless terminal of the transmission destination. The status takes two states, ACTION and FREEZE. A status of ACTION indicates that TCP communication is normally performed, and a status of FREEZE indicates that the TCP communication is frozen.

  The example of the socket information management table 20 indicates that TCP communication is being performed between the port 1 of the wireless terminal 1 and the port 2 of the wireless terminal 3 and is normally operating (ACTION). Yes. Further, TCP communication is performed between the port 2 of the wireless terminal 1 and the port 1 of the wireless terminal 7, indicating that it is normally operating (ACTION).

  The IP module 19 of the TCP communication source terminal registers the socket information by a two-step procedure. That is, the IP module 19 of the TCP communication source terminal performs socket information extraction processing when transmitting an IP packet, and then performs socket information registration processing when receiving an IP packet, thereby registering socket information. .

  FIG. 8 is a flowchart showing socket information extraction processing. With reference to the figure, when the TCP module is included in the created IP packet, the IP module 19 of the TCP communication transmission source terminal acquires the TCP header (S201).

  The IP module 19 of the transmission source terminal of the TCP communication, when SYN is set in the Code Bit of the acquired TCP header (that is, when the TCP connection requesting terminal transmits the first packet), or the acquired TCP header When SYN and ACK are set in the Code Bit (that is, when the TCP connection receiving terminal transmits the second packet), an IP header is acquired from the created IP packet (S202, S203).

  Then, the IP module 19 of the TCP communication source terminal acquires the source port number and the destination port number from the acquired TCP header when the source address in the acquired IP header matches the address of the own terminal. (S204, S205).

  Next, the IP module 19 of the TCP communication transmission source terminal acquires the transmission destination address from the acquired IP header (S206).

  Next, the IP module 19 of the transmission source terminal of TCP communication holds the acquired transmission destination address, transmission source port number, and transmission destination port number (transmission source information, transmission destination information) in the memory (S207).

  Next, the IP module 19 of the TCP communication source terminal sets a flag SYNC indicating that connection establishment processing is being performed to “1”.

  Next, the IP module 19 transmits an IP packet (S209).

  FIG. 9 is a flowchart showing socket information registration processing. With reference to the figure, when the SYNC flag is set to “1” and the received IP packet includes a TCP header, the IP module 19 of the TCP communication source terminal determines that TCP header. Is acquired (S601, S602).

  Next, the IP module 19 of the TCP communication transmission source terminal is set when SYN and ACK are set in the Code Bit of the acquired TCP header (that is, when the TCP connection requesting terminal receives the second packet), or When ACK is set in the Code Bit of the acquired TCP header (that is, when the TCP connection receiving terminal receives the third packet), the IP header is acquired from the received IP packet (S603, S604).

  Next, when the transmission destination address in the acquired IP header matches the address of the own terminal, the IP module 19 of the transmission source terminal of TCP communication stores the transmission destination address, the transmission source port number, and the transmission destination port in the memory. When the numbers (source information, destination information) are held, they are taken out from the memory, and when the socket information management table 20 is not yet created in the Internet layer, a new socket information management table 20 is created In one entry of the table, or when the socket information management table 20 has already been created, the extracted transmission destination address, transmission source port number, transmission destination port number, and status (ACTION) are added to this table. Socket information consisting of the following is written (S606, S607).

  Next, the IP module 19 of the TCP communication transmission source terminal deletes the transmission destination address, transmission source port number, and transmission destination port number written in the socket information management table 20 from the memory, and sets the SYNC flag to “0”. (S608, S609).

(4) TCP Communication Freeze Migration Control As described in (1) above, when the DSDV daemon 16 receives a link disconnection notification, it updates the metric of the path information about the link disconnection to BIG. The DSDV daemon 16 monitors the change in the metric of the route information in the routing table 21. When the metric of the route information changes to BIG, the DSDV daemon 16 performs a TCP communication freeze transition control specifying the transmission source port number.

  Here, the TCP communication freeze means that the execution is temporarily stopped while maintaining the current state of the TCP communication using the designated transmission source port number by the TCP module 17. Specifically, at the time of TCP communication freeze, the transmission of new packets is stopped, the retransmission timer is stopped, the window size immediately before the TCP communication freeze is held, and / or the average RTT (Round Trip Time) value is calculated. Absent. Hereinafter, these will be described.

  In TCP communication, packet communication is performed using a sequence number. On the transmission side, when the sequence number of the next packet to be transmitted is equal to or less than the maximum sequence number that can be transmitted, the packet can be transmitted.

  Therefore, when the TCP communication is frozen, the transmission of a new packet is stopped by setting the already received sequence number as the maximum sequence number that can be transmitted. In addition, the maximum sequence number that can be transmitted before the setting is held in another area, and when the freeze of the TCP communication is canceled, by using this stored maximum sequence number that can be transmitted, It is possible to return to the state immediately before the TCP communication freeze, and to prevent a decrease in the throughput of the TCP communication.

  In TCP communication, when a packet is transmitted, a retransmission timer is started. When an acknowledgment packet (ACK) is not returned by the time the timeout value is reached, the packet is retransmitted. This timeout value is a function of the average RTT value. Become. The average RTT value is obtained by measuring an RTT (Round Trip Time) value, which is a time from when a packet is transmitted until an acknowledgment packet (ACK) is returned, and calculating the average.

  In TCP communication, the amount of packet transmission is controlled by a sliding window method. The window size is the number of packets transmitted at a time. Each time one packet is received at the receiving side, an acknowledgment packet (ACK) is transmitted. When the transmission side receives this acknowledgment packet (ACK), it slides the window and transmits the next packet.

  In TCP communication, the window size is adjusted by a slow start and congestion avoidance algorithm. In the slow start, when the retransmission timer reaches the timeout value, the transmission side decreases the window size to 1, and then increases the window size exponentially. In congestion avoidance, the window size is increased by one when the window size becomes half the size at the time of timeout.

  Therefore, when the TCP communication freezes, it is possible to prevent unnecessary retransmission of the packet by stopping the retransmission timer.

  Further, if the window size is updated to a random value during the TCP communication freeze and the window size is started from the updated random value or the initial value after the release of the TCP communication freeze, the throughput of the TCP communication decreases. Therefore, at the time of TCP communication freeze, the window size immediately before the TCP communication freeze is retained, and after the TCP communication freeze is released, the optimum window size is set based on the retained window size, thereby reducing the TCP communication throughput. Can be prevented.

  Furthermore, when the TCP communication freeze is performed, the calculation of the average RTT value is not performed, so that unnecessary measurement of the RTT value can be prevented.

  The procedure of the TCP communication freeze migration control process by the DSDV daemon 16 will be described with reference to the flowchart shown in FIG.

  The DSDV daemon 16 monitors the value of Metric in the routing table 21, and when route information in which Metric is BIG occurs as shown in FIG. 11, the destination address of the route information from the routing table 21. Is acquired (S301, S302).

  The DSDV daemon 16 checks whether or not the socket information management table 20 exists in the Internet layer and socket information of the same destination address as the acquired destination address exists in the socket information management table 20 (S303). , S304).

  When such socket information exists, the DSDV daemon 16 checks whether or not the status of the socket information is ACTION, and when it is ACTION, the DSDV daemon 16 sends to the TCP module 17 the transmission source included in the socket information. An instruction to freeze TCP communication using the port number is given (S304, S305).

  Further, the DSDV daemon 16 updates the status of the socket information in the routing table 21 to FREEZE. Thereby, the socket information management table 20 of FIG. 7 is changed to the socket information management table 20 as shown in FIG. 12 (S306).

(5) TCP communication freeze release control When a new route to the destination terminal of TCP communication frozen in the TCP communication is discovered by the periodic search processing by the DSDV daemon 16, the Metric of the routing table 21 is changed from BIG. Updated to a value other than BIG. The DSDV daemon 16 monitors the change in Metric of the route information in the routing table 21 and performs TCP communication freeze release control when Metric changes from BIG to a value other than BIG.

  Here, the release of the TCP communication freeze is to release the freeze state of the TCP communication using the designated transmission source port number by the TCP module 17, more specifically, the transmission of a new packet is resumed and retransmitted. It means that the timer is operated, the window size is started from the value immediately before the TCP communication freeze, and / or the calculation of the average RTT value is restarted (all the packets before the restart are excluded from the calculation).

  The procedure of the TCP communication freeze release control process by the DSDV daemon 16 will be described with reference to the flowchart shown in FIG.

  The DSDV daemon 16 monitors the value of Metric in the routing table 21, and when route information in which Metric has changed from BIG to a value other than BIG is generated as shown in the routing table of FIG. Is acquired (S401, S402).

  The DSDV daemon 16 checks whether or not the socket information management table 20 exists in the Internet layer and socket information of the same destination address as the acquired destination address exists in the socket information management table 20 (S403). , S404).

  When such socket information exists, the DSDV daemon 16 checks whether or not the status of the socket information is FREEZE. An instruction to cancel the TCP communication freeze is issued (S405, S406).

  Furthermore, the DSDV daemon 16 updates the status of the socket information in the routing table 21 to ACTION. As a result, the socket information management table 20 in FIG. 12 is returned to the socket information management table 20 in FIG. 7 (S407).

(6) Deletion of socket information The IP module 19 of the source terminal of TCP communication relates to the TCP communication based on the TCP header shown in FIG. 2 and the IP header shown in FIG. 3 when the TCP communication connection is disconnected. The socket information is deleted from the socket information management table 20.

  FIG. 15 is a flowchart showing socket information deletion processing. With reference to the figure, when the TCP module is included in the created IP packet, the IP module 19 of the TCP communication transmission source terminal acquires the TCP header (S501).

  When the FIN is set in the Code Bit of the acquired TCP header (that is, when the TCP disconnection requesting terminal transmits the first packet, or the TCP communication disconnection receiving terminal When transmitting the third packet), an IP header is acquired from the created IP packet (S502, S503).

  Then, the IP module 19 of the TCP communication source terminal acquires the source port number and the destination port number from the acquired TCP header when the source address in the acquired IP header matches the address of the own terminal. (S504, S505).

  Next, the IP module 19 of the source terminal of TCP communication acquires the destination address from the acquired IP header (S506).

  Next, when the socket information including the acquired transmission destination address, transmission source port number, and transmission destination port number is registered in the socket information management table 20 of the Internet layer, the IP module 19 of the transmission source terminal of TCP communication is registered. The socket information is deleted (S507).

(Example of communication operation by ad hoc network)
Next, an example of a communication operation using an ad hoc network will be described with reference to FIG. In this example, it is assumed that the wireless terminal 1 is a TCP communication source terminal, the wireless terminal 1 is a TCP communication connection requesting terminal, and the wireless terminal 4 is a TCP communication connection accepting terminal.

  The TCP module 17 of the wireless terminal 1 and the TCP module 17 of the wireless terminal 4 perform TCP communication. When the TCP module 17 establishes a TCP communication connection, the IP module 19 of the wireless terminal 1 registers socket information in the socket information management table 20 in the Internet layer based on the TCP header and the IP header. In the socket information of the wireless terminal 1, as shown in FIG. 16, the transmission destination address is the address of the wireless terminal 4, the transmission source port number is 1, the transmission destination port number is 2, and the status is ACTION.

  The routing table 21 of the wireless terminal 1 has route information in which the transmission destination address is the address of the wireless terminal 4, the NextHop address is the address of the wireless terminal 2, the Metric is 3, and the SeqNum is 0. Further, the routing table 21 of the wireless terminal 2 has route information in which the transmission destination address is the address of the wireless terminal 4, the NextHop address is the address of the wireless terminal 3, the Metric is 2 and the SeqNum is 0. Further, the routing table 21 of the wireless terminal 3 has route information in which the transmission destination address is the address of the wireless terminal 4, the NextHop address is the address of the wireless terminal 4, Metric is 1 and SeqNum is 0. Therefore, the packet transmitted from the wireless terminal 1 is delivered to the wireless terminal 4 via the wireless terminal 2 and the wireless terminal 3.

  When the MAC module 23 of the wireless terminal 3 detects that the link is disconnected when the number of packet retransmissions exceeds a predetermined number, the DSDV daemon detects that the link has been disconnected and the IP packet that has not been transferred. 16 is notified. The DSDV daemon 16 of the wireless terminal 3 has the address of the wireless terminal 4 that is the transmission destination address of the IP header of the IP packet that has not been transferred, in the routing table 21. The metric of the route information that is the address of the wireless terminal 4 is updated to BIG. The DSDV daemon 16 of the wireless terminal 3 creates an Update packet including the changed route information, and sends the Update packet to the UDP module 18 for broadcast transmission.

  The DSDV daemon 16 of the wireless terminal 2 receives the Update packet by the UDP module 18. The DSDV daemon 16 of the wireless terminal 2 transmits the transmission in the routing table 21 because the address of the wireless terminal 4 that is the transmission destination address of the route information included in the Update packet exists in the routing table 21 and its Metric is not BIG. The metric of the route information whose destination address is the address of the wireless terminal 4 is updated to BIG. The DSDV daemon 16 of the wireless terminal 2 creates an Update packet including the changed route information, and sends the Update packet to the UDP module 18 for broadcast transmission.

  The DSDV daemon 16 of the wireless terminal 1 receives the Update packet by the UDP module 18. The DSDV daemon 16 of the wireless terminal 1 transmits the transmission in the routing table 21 because the address of the wireless terminal 4 that is the transmission destination address of the route information included in the Update packet exists in the routing table 21 and its Metric is not BIG. The metric of the route information whose destination address is the address of the wireless terminal 4 is updated to BIG.

  The DSDV daemon 16 of the wireless terminal 1 monitors the value of Metric in the routing table 21, detects the occurrence of route information in which Metric becomes BIG, and transmits the wireless address that is the destination address of the route information from the routing table 21. The address of the terminal 4 is acquired. The DSDV daemon 16 of the wireless terminal 1 has a socket information management table 20 in the Internet layer, and socket information with the acquired address of the wireless terminal 4 as the transmission destination address in the socket information management table 20. It is checked whether or not the status of the socket information is ACTION. Since the status is ACTION, the DSDV daemon 16 of the wireless terminal 1 instructs the TCP module 17 to freeze TCP communication using the transmission source port number included in the socket information. The TCP module 17 freezes the TCP communication using the instructed transmission source port number. Further, the DSDV daemon 16 of the wireless terminal 1 updates the status of the socket information in the routing table 21 to FREEZE.

  As described above, according to the communication method in the ad hoc network that performs routing using the table-driven routing protocol according to the embodiment of the present invention, the TCP communication is frozen when the link is disconnected, and a new link is created. By releasing the TCP communication freeze when is connected, it is possible to appropriately control the TCP communication and prevent a decrease in the throughput of the TCP communication.

  The present invention is not limited to the above embodiment, and includes, for example, the following modifications.

(1) Link disconnection In the embodiment of the present invention, the MAC module 23 detects that the link is disconnected when the number of packet retransmissions exceeds a predetermined number. However, the present invention is not limited to this. . For example, when the ARP buffer stored when the IP address and the MAC address cannot be matched overflows, the link disconnection may be detected.

(2) Table-driven routing protocol In the embodiment of the present invention, the DSDV protocol is taken as an example of the table-driven routing protocol. However, the present invention is not limited to this, and other table-driven routing protocols are used. However, it can be carried out in the same manner.

(3) Metric indicating link disconnection
In the embodiment of the present invention, the link is disconnected when Metric is a predetermined value BIG. However, the present invention is not limited to this, and the link is disconnected when Metric is a predetermined symbol. It is also possible to express that

(4) DSDV daemon In the embodiment of the present invention, the DSDV daemon (1) ad hoc network routing, (2) link disconnection detection and notification, (4) TCP communication freeze transition control, (5) TCP communication Although freeze release control is performed, the present invention is not limited to this. For example, (1) and (2) may be performed by the DSDV daemon, and (4) and (5) may be performed by one or two daemons other than the DSDV daemon.

(5) Application and Process / Application Protocol In the embodiment of the present invention, description has been given using Mailer as an application and SMTP as a process / application protocol. However, the present invention is not limited to this, and TCP is used as a transport layer. As long as it uses, any thing may be used.

(6) TCP communication freeze In place of or in addition to the TCP communication freeze processing described in the embodiment of the present invention, for example, the following processing may be performed.

(A) ACK reception process in TCP communication in freeze state Even if the TCP communication on the transmission side is in the freeze state, the TCP communication on the reception side cannot know it. For this reason, even if link disconnection occurs on the path from the transmission side, the reception side does not matter, and ACK is transmitted. As a result, the ACK may be received even though the TCP communication on the transmission side is in a frozen state. In order to cope with this, the following processing is performed.

  The transmitting TCP in the frozen state does not transmit a new packet even if it receives an ACK. The sending TCP in the frozen state compares the sequence number registered in the received ACK with the sequence number registered in the ACK received immediately before. As a result, when it is determined that a new ACK has been received, the transmission side TCP updates the sequence number of the ACK received immediately before held by the transmission side TCP and the maximum sequence number received. When an ACK having a sequence number that overlaps with the sequence number of the ACK received immediately before is received, only the number of times of duplication is counted, and no other reception processing is performed.

(B) Queue control in DSDV notified of link disconnection In DSDV, when a route is not found, a packet to be transmitted or transferred is stored in a queue. Packet loss during link disconnection is avoided by preferentially transmitting or transferring packets stored in the queue after route recovery.

  However, a threshold is also provided in the DSDV queue, and when a packet exceeding the threshold is sent, the oldest packet stored in the queue is discarded by the FIFO. Further, by storing the packet in the DSDV queue, even if the TCP parameter is controlled after the path recovery, the packet that cannot be controlled from the TCP is transmitted from the DSDV.

  Therefore, all the packets received by DSDV during link disconnection are discarded, and after the TCP communication freeze is canceled, an inquiry is made from the sending TCP to the receiving TCP, and all packets discarded by DSDV are also discarded. When TCP performs retransmission, unified management of packet transmission by TCP is possible, and communication control can be facilitated.

(7) Deletion of socket information In the embodiment of the present invention, the IP module 17 of the TCP communication disconnection request terminal deletes the socket information when transmitting the first packet, but receives the second and third packets. In addition, the socket information may be deleted. The IP module 17 of the TCP communication disconnection receiving terminal deletes the socket information when transmitting the second and third packets, but may delete the socket information when receiving the fourth packet.

(8) Reference to Code Bit In S203, when the TCP connection requesting terminal transmits the first packet when the IP module 19 of the TCP communication transmission source terminal is set to SYN in the Code Bit of the acquired TCP header It is identified that the TCP connection receiving terminal transmits the second packet when SYN and ACK are set in the Code Bit of the acquired TCP header. However, the present invention is not limited to this. . The IP module 19 of the TCP communication transmission source terminal transmits the first packet when the TCP connection requesting terminal transmits the first packet when SYN is set in the Code Bit of the acquired TCP header, or the TCP connection receiving terminal transmits the second packet. It is good also as what identifies as it is time to transmit. That is, the transmission of the second packet may be determined with reference to only SYN.

  In step S603, the IP module 19 of the TCP communication transmission source terminal identifies that the TCP connection requesting terminal has received the second packet when SYN and ACK are set in the Code Bit of the acquired TCP header. Although it has been identified that the time when ACK is set in the Code Bit of the acquired TCP header is the time when the TCP connection receiving terminal has received the third packet, the present invention is not limited to this. The IP module 19 of the TCP communication transmission source terminal identifies that the TCP connection requesting terminal has received the second packet when SYN is set in the Code Bit of the acquired TCP header, and acquires the acquired TCP header. It is also possible to identify that the time when the ACK is set in the Code Bit is that the TCP connection receiving terminal has received the third packet. That is, the reception of the second packet may be determined with reference to only SYN.

  The embodiment disclosed this time should be considered as illustrative in all points and not restrictive. The scope of the present invention is defined by the terms of the claims, rather than the description above, and is intended to include any modifications within the scope and meaning equivalent to the terms of the claims.

It is a figure which shows the structure of an ad hoc network. 1 is a diagram illustrating a configuration of a wireless terminal 1. FIG. It is a figure which shows the structure of an IP header. It is a figure which shows the structure of a TCP header. It is a figure which shows the example of the routing table. 6 is a flowchart illustrating a procedure of link disconnection detection and notification processing by the DSDV daemon 16; It is a figure which shows the example of the socket information management table 20 when the transmission origin terminal of TCP communication is the wireless terminal 1. It is a flowchart which shows the procedure of the extraction process of the socket information by IP module 19. It is a flowchart which shows the procedure of the registration process of the socket information by IP module 19. 6 is a flowchart showing a procedure of TCP communication freeze transition control processing by the DSDV daemon 16; It is a figure which shows the example of the routing table. It is a figure which shows the example of the socket information management table 20 when the transmission origin terminal of TCP communication is the wireless terminal 1. 6 is a flowchart showing a procedure of TCP communication freeze release control processing by the DSDV daemon 16; It is a figure which shows the example of the routing table. 10 is a flowchart showing a procedure of socket information deletion processing by the DSDV daemon 16; It is a figure which shows the example of communication operation | movement of the communication system of the ad hoc network which concerns on embodiment of this invention.

Explanation of symbols

  1 to 9 wireless terminal, 11 input unit, 12 display unit, 13 e-mail application, 14 communication control unit, 15 SMTP module, 16 DSDV daemon, 17 TCP module, 18 UDP module, 19 IP module, 20 socket information management table, 21 routing table, 22 LLC module, 23 MAC module, 24 radio interface module, 25 antenna.

Claims (8)

A communication method in an ad hoc network that performs routing using a table-driven routing protocol,
A TCP communication source terminal, when establishing a TCP communication connection, registering socket information including an address of the TCP communication destination terminal, a source port number, and a destination port number in a socket information management table; ,
When each terminal detects link disconnection, the corresponding route information metric in the routing table of its own terminal is updated to a value or symbol indicating link disconnection, and a packet including the updated route information is broadcasted When,
When each terminal receives a packet containing the updated route information, the corresponding route information metric in the routing table of its own terminal is updated to a value or symbol indicating link disconnection, and the updated route information is included. Broadcasting a packet;
The source terminal of the TCP communication acquires the address of the destination terminal of the route information from the routing table when the metric of the route information in the routing table of the terminal becomes a value or symbol indicating link disconnection. And referring to the socket information management table, identifying socket information including an address of the acquired destination terminal, and freezing TCP communication using a source port number included in the identified socket information; A communication method in an ad hoc network. The socket information further includes a status indicating whether the TCP communication state is frozen or an action indicating a normal operation state,
In the step of registering the socket information, the source terminal of the TCP communication sets the socket information further including a status indicating an action in addition to the address of the destination terminal, the source port number, and the destination port number. Register in the information management table,
In the step of freezing the TCP communication, the source terminal of the TCP communication freezes the TCP communication using the source port number of the socket information only when the status of the socket information is an action,
The communication method in the ad hoc network according to claim 1, further comprising the step of the TCP communication source terminal updating the status of the socket information to freeze after the TCP communication freeze. The communication method in the ad hoc network further includes:
When each terminal discovers a new route, the routing table is updated with the route information of the new route with the metric as the number of hops, and the updated route information is broadcasted.
When the metric of the route information in the routing table of the terminal of the TCP communication changes from a value or symbol indicating link disconnection to the number of hops, the address of the destination terminal of the route information from the routing table The socket information management table is referenced to identify socket information including the address of the acquired destination terminal, and only when the status of the identified socket information is frozen, Canceling the freeze state of TCP communication using the source port number;
The communication method in the ad hoc network according to claim 2, further comprising a step of updating a status of the socket information to an action after the TCP communication source terminal is released from the freeze state of the TCP communication.   The communication method in an ad hoc network according to claim 3, wherein the table-driven routing protocol is a DSDV protocol.   The TCP communication source terminal deletes, from the socket information management table, socket information including the TCP communication destination terminal address, source port number, and destination port number when the TCP communication connection is disconnected. The communication method in the ad hoc network according to claim 1, further comprising: A communication method in an ad hoc network that performs routing using a table-driven routing protocol,
When the IP module that executes the IP of the TCP communication source terminal establishes a TCP communication connection by the TCP module that executes the TCP, the address of the TCP communication destination terminal, the source port number, and the destination Registering socket information including the port number in the socket information management table of the Internet layer;
When the daemon that executes the table-driven routing protocol of each terminal receives notification from the MAC module that executes the MAC protocol that the link has been disconnected together with the packet that has not been transferred, Updating a corresponding route information metric in the routing table to a value or symbol indicating link disconnection, and broadcasting a packet including the updated route information by a UDP module that executes the UDP protocol;
When the daemon of each terminal receives a packet including the updated route information, the metric of the corresponding route information in the routing table of its own terminal is updated to a value or symbol indicating link disconnection, and the updated route Broadcasting a packet containing information by the UDP module;
The daemon of the TCP communication source terminal monitors the metric of the route information in the routing table of its own terminal, and when the metric of the route information becomes a value or symbol indicating link disconnection, The address of the destination terminal of the route information is acquired, the socket information including the address of the acquired destination terminal is specified with reference to the socket information management table, and the specified socket information is specified for the TCP module. A communication method in an ad hoc network, including the step of freezing TCP communication using a transmission source port number included in. The step of registering the socket information includes:
When the IP module of the TCP communication source terminal transmits an IP packet, obtaining a TCP header from the IP packet;
When the IP module of the TCP communication source terminal is set to SYN in the Code Bit in the TCP header, the step of acquiring an IP header from the IP packet;
When the IP module of the source terminal of the TCP communication is the address of the source terminal in the IP header, the source port number and the destination port number are acquired from the acquired TCP header, Obtaining a destination terminal address from the obtained IP header, and holding the obtained destination terminal address, source port number, and destination port number in a memory;
A step of setting a flag indicating that the IP module of the TCP communication source terminal is in the process of establishing a connection to a predetermined value;
When the IP module of the TCP communication source terminal receives an IP packet when the flag is set to the predetermined value, obtaining a TCP header from the IP packet;
When the IP module of the source terminal of the TCP communication is set to SYN or ACK in the Code Bit in the TCP header, a step of acquiring an IP header from the IP packet;
The IP module of the source terminal of the TCP communication, when the address of the destination terminal in the IP header is the address of the own terminal, the address of the destination terminal, the source port number and the transmission held in the memory The communication method in the ad hoc network according to claim 6, further comprising: registering socket information including a destination port number and an action indicating a normal operation state as a status in a socket information management table. The communication method in the ad hoc network further includes:
When the IP module of the TCP communication source terminal transmits an IP packet, obtaining a TCP header from the IP packet;
When the IP module of the TCP communication source terminal has FIN set in the Code Bit in the TCP header, the IP module further obtains an IP header from the IP packet;
When the IP module of the source terminal of the TCP communication is the address of the source terminal in the IP header, the source port number and the destination port number are acquired from the acquired TCP header, Obtaining a destination terminal address from the obtained IP header;
The IP module of the source terminal of the TCP communication includes a step of deleting socket information including the acquired destination terminal address, source port number, and destination port number from the socket information management table. Item 8. A communication method in an ad hoc network according to Item 7.

Images