JP4741964B2 - COMMUNICATION DEVICE, COMMUNICATION SYSTEM, AND COMMUNICATION METHOD - Google Patents

Description

  The present invention relates to a communication device, a communication system, and a communication method, and more particularly, to a communication device, a communication system, and a communication method that can establish a P2P communication path between two different private networks.

  Generally, in order to identify a communication terminal connected to the network, an IP address is assigned to the communication terminal. An IP address that is uniquely determined in the entire connected network and enables one-to-one communication (peer-to-peer communication: hereinafter referred to as “P2P communication”) is referred to as a global IP address. An IP address that is uniquely determined only within a specific network and cannot be directly used for P2P communication with another network is referred to as a local IP address. A network in which communication terminals are mutually identified by a local IP address is called a private network.

  In order for a communication terminal (communication terminal having a local IP address) connected to a private network to communicate with a communication device (communication terminal having a global IP address) connected to the global network, generally, the private network and the global network A communication device (router) having a network address translation (Network Address Translation: hereinafter referred to as “NAT”) function or a network address port translation (Network Address Port Translation: hereinafter referred to as “NAPT”) function. Is connected. In the following, NAT and NAPT are collectively referred to as “NAT” for simplification of description.

  However, by simply relaying between the private network and the global network by a communication device such as a router by the NAT function of the router, two communication devices connected to different private networks can be used for P2P communication. Packets cannot be exchanged.

  Therefore, STUN (Simple Traversal UDP through NAT: RFC3389) is used as a method for two terminals, each connected to a different private network via a router, to establish a P2P communication path beyond the NAT of the router. How to do is known. In the following description, “packet” means “UDP packet”.

  First, prior to the description of the P2P communication path establishment method using STUN, the type of NAT will be described with reference to FIGS. 18A to 18E. The private network is not limited to being constructed as a home, that is, a home network, but may be constructed as a corporate network. In the following, for convenience of explanation, it is assumed that the private network is a home network, but the following explanation is similarly applied to a large-scale private network constructed in a company. In addition, when viewed from a private network in a certain home, a network different from the home network is referred to as an external network or an external network. Further, for convenience of explanation, the combination of the IP address IP and the port number P is expressed as (IP, P).

  FIG. 18A is a diagram for explaining NAT called Full Cone NAT.

  A router having the Full Cone NAT characteristic always assigns the same port number to a packet transmitted from a specific port number of a home terminal having a certain private address to a certain outside terminal. For example, the router always assigns the port number [Pb] to the packet transmitted by the home terminal A (IPa, Pa), and transfers the packet to the external terminal C connected to the external network. In addition, the router having the Full Cone NAT characteristic transfers all packets transmitted from the outside terminal connected to the external network to (IPb, Pb) to the inside terminal (IPa, Pa). Therefore, the router also transfers a packet transmitted from the out-of-home terminal D (IPd, Pd) to (IPb, Pb) to the in-home terminal A (IPa, Pa).

  FIG. 18B is a diagram for explaining a NAT called Restricted Cone NAT (hereinafter referred to as “R NAT”).

  The router having the R NAT characteristic always assigns the same port number [Pb] to a packet transmitted from a certain home terminal A (IPa, Pa) to the outside network. However, the router having the R NAT characteristic allows only the packet transmitted from the remote terminal C (having the IP address IPc), which is the transmission destination of the packet transmitted from the remote terminal A, to (IPb, Pb). It is different from the router having the Full Cone NAT characteristic in that it transfers to (IPa, Pa). Therefore, the router does not transfer the packet transmitted to (IPb, Pb), which is not the destination of the packet transmitted by the home terminal A, to the home terminal A. Note that the router having the R NAT characteristic transfers packets of all the source port numbers to the home terminal A if the source IP address of the received packet is [IPc]. Therefore, the router also transfers the packet transmitted from the port [Pc2] of the outside terminal C to (IPb, Pb) to the in-home terminal A.

  FIG. 18C is a diagram for explaining a NAT characteristic called Port Restricted Cone NAT (hereinafter referred to as “PR NAT”).

  The router having the PR NAT characteristic always assigns the same port number [Pb] to a packet transmitted from a certain home terminal A (IPa, Pa) to the outside network. However, the router having the PR NAT characteristic transfers only the packet transmitted from the destination (IPc1, Pc1) of the packet transmitted by the home terminal A to (IPb, Pb) to the home terminal A (IPa, Pa). In this respect, it differs from a router having R NAT characteristics. Therefore, the router discards not only the packet transmitted from the source (IPd, Pd) of the remote terminal D but also the packet transmitted from the remote source (IPc2, Pc2).

  Hereinafter, the above Full Cone NAT, R NAT, and PR NAT are collectively referred to as “Cone NAT”.

  FIG. 18D is a diagram for explaining a NAT called Address Sensitive Symmetric NAT (hereinafter referred to as “AS NAT”).

  A router having AS NAT characteristics assigns a different port number to each destination IP address of a packet transmitted from the home terminal A. For example, a port number [Pb1] is assigned to a packet transmitted from the home terminal A (IPa, Pa) to the remote terminal C (IPc, Pc), and the remote terminal is assigned from the home terminal A (IPa, Pa). A port number [Pb2] is assigned to a packet transmitted to an out-of-home terminal D (IPd, Pd) having an IP address different from. A router having AS NAT characteristics transfers only packets transmitted from an out-of-home terminal assigned to a certain port number to the in-home terminal A (IPa, Pa), and the out-of-home terminal not assigned to the port number Discard all packets sent from. For example, the router transfers a packet transmitted to (IPb, Pb) from the out-of-home terminal C having the IP address [IPc] to the in-home terminal A (IPa, Pa). The packet transmitted to the (IPb, Pb) addressed from the remote terminal D is discarded. Note that a router having R NAT characteristics forwards packets having all source port numbers to the home terminal A if the source IP address of the packet transmitted to (IPb, Pb) is [IPc]. To do.

  FIG. 18E is a diagram for describing NAT called Port Sensitive Symmetric NAT (hereinafter referred to as “PS NAT”).

  The router having the PS NAT characteristic assigns a different port number to each combination of the destination IP address and the destination port number of the packet transmitted by the home terminal A, and transmits the received packet to the network. For example, the router assigns the port number [Pb0] to the destination (IPc, Pc0) of the packet transmitted from the home terminal A (IPa, Pa) and the port number [Pb1] to the destination (IPc, Pc1). And port number [Pb2] is assigned to the destination (IPc, Pc2). Further, the router transfers a packet transmitted from an out-of-home terminal assigned to a certain port number to the in-home terminal A, but discards all packets transmitted from other than the out-of-home terminal assigned to the port number. Therefore, the router transfers the packet transmitted from the source (IPc, Pc1) with the port number [Pb1] as the destination to the home terminal A, but the source (IPc, Pc2) and the source (IPd, Pd1) All the packets transmitted from to the port number [Pb1] are discarded.

  Hereinafter, the above AS NAT and PS NAT are collectively referred to as “Symmetric NAT”.

  In addition to the above-described Cone NAT and Symmetric NAT, a router having a property of setting NAT using the same port number as the source port number of the home terminal is known.

  FIG. 19A is a diagram for explaining the outline of the Port Reuse characteristic.

  As shown in FIG. 19A, the router assigns a port number [Pa] to a packet whose source is the port number [Pa] of the home terminal A, and a packet whose source is the port number [Pb] of the home terminal A. Is assigned a port number [Pb], and a port number [Pc] is assigned to a packet originating from the port number [Pc] of the home terminal A. Thus, the characteristic that the communication device assigns the same port number as the source port number to a packet including a source port number is called “Port Reuse”.

  Port Reuse is a characteristic different from the NAT characteristic described above. Therefore, a communication device having the Port Reuse characteristic always has one of the NAT characteristics described above.

  FIG. 19B is a diagram for explaining the function of the router having the Port Reuse characteristic.

  When the router shown in FIG. 19B has the “Port Reuse characteristic and the Full Cone NAT characteristic”, the router always sets the port to (IPa, Pa) for the source IP address and the source port number. The number [Pa] is assigned and the packet is sent to the external network. In addition, the router transfers to the home terminal A all the packets transmitted from the outside terminal connected to the external network to the destination (IPb, Pa).

  Next, a method for using STUN to establish a P2P communication path beyond the NAT of the router will be described.

  In order to establish a P2P communication path between two communication terminals each connected to a different private network via a router, each terminal may set a destination IP address and a destination port number in the packet. is necessary. However, as described above, since the address translation method differs depending on the NAT characteristic of the router, the communication terminal needs to set the destination of the packet in consideration of the NAT characteristic of the router.

  Therefore, in order to establish the P2P communication path, first, the NAT characteristic of the router to which the communication terminal is connected is investigated according to the method specified by STUN. More specifically, in STUN, a communication terminal transmits a test packet to each of two servers having different IP addresses. Then, by determining whether or not the port number assigned to each of the test packets by the router matches, it is determined whether the NAT characteristic of the router is Cone NAT or Symmetric NAT.

  Hereinafter, a method for investigating NAT characteristics by STUN will be described with reference to FIGS. 20A to 20B and FIG.

  FIG. 20A is a diagram illustrating an outline of STUN-TestI.

  In Test I, the home terminal A transmits a packet whose source is (IPa, Pa) to the server 1 (IPc, Pc1). The server 1 refers to the packet transmitted from the home terminal A and transmits a packet in which [Pc1] is set as the source port number and [Pb] assigned by the router is set as the destination port number. .

  FIG. 20B is a diagram showing an outline of STUN-Test II.

  In Test II, the home terminal A transmits a packet whose source is (IPa, Pa) and whose destination is (IPc, Pc1) of the server 1. In response to this, the server 2 transmits a packet in which (IPd, Pd1) is set as the source and [Pb] assigned by the router is set as the destination port number.

  FIG. 20C is a diagram showing an outline of STUN-TestIII.

  In Test III, the home terminal A transmits a packet whose source is (IPa, Pa) and whose destination is (IPc, Pc1) of the server 1. The server 1 transmits a packet in which [Pc2] different from [Pc1] is set as the source port number and the port number [Pb] is set as the destination port number.

  FIG. 21 is a flowchart for investigating NAT characteristics by STUN.

  First, the home terminal A executes the above Test I (step S2501), and determines whether a response is received from the server 1 (step S2502). If the home terminal A receives a response from the server 1, the process proceeds to step S2504. Otherwise, the home terminal A determines that UDP communication with the server 1 is not possible (step S2503). .

  When the home terminal A receives the response from the server 1 (Yes in step S2501), the home terminal A sends the IP address [IPb] after the address conversion included in the received response packet and its own IP address [IPa]. Are matched (step S2504). The home terminal A proceeds to step S2505 if the IP address [IPb] included in the response packet matches its own IP address [IPa], and proceeds to step S2909 otherwise. The coincidence of the IP addresses [IPb] and [IPa] indicates that the home terminal A is connected to the external network without going through the router, and NAT is not performed between the home terminal A and the server 1. .

  If Yes in step S2504, home terminal A executes Test II (step S2505) and determines whether a response from server 2 has been received (step S2506). When the home terminal A receives a response from the server 2 (Yes in step S2506), the home terminal A can receive a packet from a communication terminal other than the destination terminal (server 1). It is determined that it is open (step S2508). On the other hand, if the home terminal A does not receive a response from the server 2 (No in step S2506), the home terminal A determines that packet transmission / reception is restricted by the UDP firewall (step S2507).

  If the IP addresses [IPb] and [IPa] do not match as a result of Test I (No in step S2504), the in-home terminal A executes Test II (step S2509) and receives from the server 2 different from the packet destination. It is determined whether a response has been received (step S2510). When the home terminal A receives the response from the server 2 (Yes in step S2510), the router to which the home terminal A is connected determines that it has the Full Cone NAT characteristic (step S2511). On the other hand, when the home terminal A does not receive a response from the server 2 (No in step S2510), the home terminal A executes Test I for the server 2 (step S2512). Here, the port number assigned by the router to the transmission packet from the home terminal A is [Pb ′]. The home terminal A includes the combination (IPb, Pb ′) of the IP address and port number set by the NAT included in the response packet received from the server 2 in step S2512 and the response packet received from the server 1 in step S2501. It is determined whether or not the combination (IPb, Pb) of the IP address and port number set by NAT matches with each other (step S2513). If the home terminal A determines that (IPb, Pb ') matches (IPb, Pb) (Yes in step S2513), the home terminal A proceeds to step S2515. If the home terminal A determines that (IPb, Pb ′) and (IPb, Pb) do not match (No in step S2513), the home terminal A determines that the router to which it is connected has Symmetric NAT characteristics (step S2513). S2514).

  If Yes in step S2513, home terminal A executes Test III (step S2515) and determines whether a response has been received from server 1 (step S2516). When the home terminal A receives a response from the server 1 (Yes in step S2516), it determines that the router to which it is connected has R NAT characteristics (step S2517). In other cases, it connects itself. It is determined that the router to be used has the PR NAT characteristic (step S2518).

  As described above, in the method defined in STUN, the NAT characteristics are investigated by executing the three types of tests shown in FIGS. 20A to 20C according to the processing flow shown in FIG.

  Also, as a result of investigating NAT characteristics by STUN, when two communication terminals are both connected to a router having the Cone NAT characteristic, a method of establishing a P2P communication path between the two communication terminals is also known. . Hereinafter, a P2P communication path establishment method using STUN will be described with reference to FIG.

  FIG. 22 is a sequence diagram showing a P2P communication path establishment method using STUN.

  First, the terminal 1 transmits an IP / port registration request packet for requesting registration of an IP address and a port number to the server via the router 1 (step S201). The router 1 converts the source of the packet received from the terminal 1 from (IPL1, LP1) to (IPG1, GP1), and transmits the address-converted packet to the server 1. The server registers the source IP address IPG1 and source port number GP1 included in the packet received from the router 2 (step S202).

  Similarly, the terminal 2 transmits an IP / port registration request packet to the server via the router 2 (step S203). The router 2 converts the source of the packet received from the terminal 2 from (IPL2, LP2) to (IPG2, GP2), and transmits the address-converted packet to the server 2. The server registers the source IP address IPG1 and source port number GP1 included in the packet received from the router 2 (step S204).

  Next, in order to perform P2P communication with the terminal 1, the terminal 2 transmits an IP / port acquisition request for accessing the terminal 1 to the server (step S205). In response to the IP / port acquisition request from the terminal 2, the server includes an IP / port including a combination (IPG1, GP1) of the IP address and port number assigned by the router 1 to the source (IPL1, LP1) of the terminal 1 An acquisition response packet is returned to the terminal 2 (step S206).

  The terminal 2 refers to the packet received from the server and acquires the IP address and port number (IPG1, GP1) for accessing the terminal 1. Therefore, the terminal 2 transmits a P2P start request packet destined for (IPG1, GP1) (step S207).

  Here, the correspondence of the router 1 to the P2P start request packet transmitted from the terminal 2 differs depending on the NAT characteristics provided in the router 1.

  First, when the router 1 executes Full Cone NAT (FIG. 18A), the router 1 transfers the P2P request packet transmitted from the terminal 2 to the terminal 1, whereby P2P communication is performed between the terminal 1 and the terminal 2. The road is established.

  Next, when the router 1 executes R NAT (FIG. 18B) or PR NAT (FIG. 18C), the router 1 discards the P2P start request packet transmitted from the terminal 2. Therefore, the P2P start request packet is not transferred to the terminal 1 (IPL1, LP1).

  It should be noted here that the router 2 having the Cone NAT characteristic transfers a packet in which (IPL2, LP2) is set as the source and (IPG1, GP1) is set as the destination to the router 1 Thus, NAT is set so that a response packet from the router 1 can be received. That is, when the router 2 executes Full Cone NAT, the router 2 forwards packets sent from all the out-of-home terminals to the router 2 (IPG2, GP2) to the terminal 2 (IPL2, LP2). Set NAT. When the router 2 has the R NAT characteristic, the router 2 sets the NAT so as to transfer a packet whose source IP address is [IPG1] to the terminal 2 (IPL2, LP2). When the router 2 has the PR NAT characteristic, the NAT is set so as to transfer a packet whose source IP address and source port number are (IPG1, GP1) to the terminal 2 (IPL2, LP2). As a result, if the router 2 receives a packet whose source IP address and source port number are (IPG1, GP1) next, the router 2 can forward the packet to the terminal 2, A P2P communication path is established between the terminal 1 and the terminal 2.

  Therefore, the terminal 1 transmits a packet for accessing the terminal 2 to the server as an IP / port acquisition request (step S208). In response to the IP / port acquisition request from the terminal 1, the server transmits a packet including the IP address and the port number (IPG2, GP2) assigned by the router 2 to the source (IPL1, LP1) of the terminal 1. It returns to the terminal 2 as a port acquisition response (step S209).

  The terminal 1 refers to the IP / port acquisition response received from the server, and uses the IP address and port number (IPG2, GP2) of the router 2 assigned to the IP address and port number (IPL2, PL2) of the terminal 2. get.

  Next, the terminal 1 transmits a P2P start request packet whose source is set to (IPL1, LP1) to the terminal 2 (step S210). As described above, at this stage, the router 2 transfers the packet whose source IP address and source port number are (IPL1, LP1) to the terminal 2 without discarding the packet. On the other hand, the router 1 can set the NAT so that the response packet from the router 2 can be received by transferring the packet in which (IPL1, LP1) is set to the source to the router 2. . That is, when the router 1 executes R NAT, the router 1 is set to transfer a packet whose source IP address is [IPG2] to the terminal 1 (IPL1, LP1). When the router 1 executes PR NAT, it is set so that a packet whose source IP address and source port number are (IPG2, GP2) is transferred to the terminal 2 (IPL2, LP2).

  Therefore, when the terminal 2 transmits a P2P start response packet to the terminal 1 in response to the P2P start request packet from the terminal 1 (step S211), a P2P communication path is established between the terminal 1 and the terminal 2.

  According to the above method, when two communication terminals connected to different private networks are both connected to a router having the Cone NAT characteristic, a P2P communication path is established between the two communication terminals. It is possible.

  However, when two communication terminals are connected via Symmetric NAT, a P2P communication path cannot be established by a technique using STUN. Hereinafter, the reason will be described.

  FIG. 23 is a sequence diagram illustrating a processing procedure when a P2P communication path establishment method using STUN is applied to two communication terminals connected via a router having the Symmetric NAT characteristic. More specifically, in the example shown in FIG. 23, a terminal 1 connected to a certain private network is connected to an external network via a router 1 that executes Symmetric NAT.

  First, in steps S301 to S304, as in steps S201 to S204 shown in FIG. 22, each of terminal 1 and terminal 2 transmits an IP / port registration request packet to the server, whereby an IP address and a port number are transmitted. Is registered on the server.

  Next, the terminal 2 transmits an IP / port acquisition request packet to the server in order to acquire an IP address and a port number used for P2P communication with the terminal 1 (step S305). In response to the IP / port acquisition request from the terminal 2, the server includes the IP address and port number (IPG1, GP1) assigned by the router 1 to the source IP address and source port number (IPL1, LP1). An IP / port acquisition response packet is returned to the terminal 2 (step S306).

  The terminal 2 refers to the packet received from the server and acquires the IP address and port number (IPG1, GP1) for accessing the terminal 1. Therefore, the terminal 2 transmits a packet destined for (IPG1, GP1) as a P2P start request (step S207).

  Here, when the router 1 has the Symmetric NAT characteristic, the router 1 discards the P2P start request packet without transferring it to the terminal 1 (IPL1, LP1). Further, as described above, the router 2 that executes Cone NAT transfers the packet in which (IPL2, LP2) is set as the source and (IPG1, GP1) is set as the destination to the router 1 Thus, NAT is set so that a response packet from the router 1 can be received.

  Next, the terminal 1 transmits a packet for accessing the terminal 2 to the server as an IP / port acquisition request (step S308). In response to the IP / port acquisition request from the terminal 1, the server assigns the IP address and port number (IPG2, GP2) assigned by the router 2 to the source IP address and source port number (IPL2, LP2) of the terminal 2 Is returned to the terminal 2 as an IP / port acquisition response (step S309).

  The terminal 1 refers to the IP / port acquisition response received from the server, and uses the IP address and port number (IPG2, GP2) of the router 2 assigned to the IP address and port number (IPL2, PL2) of the terminal 2. get.

  Next, the terminal 1 transmits, to the terminal 2 as a P2P start request, a packet whose source is (IPL1, LP1) (step S310). However, since the router 1 has a Symmetric NAT characteristic, a different port number is assigned to each destination of the packet. Therefore, the router 1 assigns a port number [GP3] different from [GP1] to the P2P start request packet received from the terminal 1, and forwards the packet to the router 2.

  Here, when the router 2 has the Full Cone NAT characteristic or the R NAT characteristic, the router 2 forwards the received P2P start request packet to the terminal 2 (IPL2, LP2). A P2P communication path is established.

  However, when the router 2 has the PR NAT or Symmetric NAT characteristic, the router 2 discards the received P2P start request packet without transferring it to the terminal 2, so that the P2P communication path between the terminal 1 and the terminal 2 Fails to establish.

  FIG. 24 is a sequence diagram showing a processing procedure when a P2P communication path establishment method using STUN is applied to two communication terminals connected via a router having the Symmetric NAT characteristic. More specifically, in the example shown in FIG. 24, a terminal 2 connected to a certain private network is connected to an external network via a router 2 that executes Symmetric NAT.

  First, in steps S401 to S404, as in steps S201 to S204 shown in FIG. 22, each of terminal 1 and terminal 2 transmits an IP / port registration request packet to the server, whereby an IP address and a port number are transmitted. Is registered on the server.

  Next, the terminal 2 transmits an IP / port acquisition request packet to the server in order to acquire an IP address and a port number for P2P communication with the terminal 1 (step S405). In response to the IP / port acquisition request from the terminal 2, the server includes the IP address and port number (IPG1, GP1) assigned by the router 1 to the source IP address and source port number (IPL1, LP1). An IP / port acquisition response packet is returned to the terminal 2 (step S406).

  The terminal 2 refers to the packet received from the server and acquires the IP address and port number (IPG1, GP1) for accessing the terminal 1. Therefore, the terminal 2 transmits a packet destined for (IPG1, GP1) as a P2P start request (step S407).

  Here, the destination of the P2P start request packet transmitted from the terminal 2 in step S407 is different from the destination of the IP / port registration request packet transmitted from the terminal 2 in step S403. Therefore, the router 2 that executes Symmetric NAT assigns a port number [GP3] different from the port number [GP2] to the P2P start request packet. That is, the router 2 sets the packet transmitted by the terminal 1 with the router 2 (IPG2, GP3) as a destination to be transferred to the terminal 2 (IPL2, LP2).

  Therefore, after obtaining the IP address and port number (IPG2, GP2) for the terminal 1 to access the terminal 2 from the server (steps S408 to S409), the destination IP address and the destination port number (IPG2, GP2) are sent to the destination. Even if the P2P start request packet is transmitted, the router 2 discards the P2P start request packet. Therefore, establishment of the P2P communication path between the terminal 1 and the terminal 2 fails.

  As described above, when at least one of the router 1 or the router 2 executes Symmetric NAT, it is difficult to say that the P2P communication path establishment method using STUN is highly likely to succeed.

  Conventionally, with respect to the above-mentioned STUN problem, a method of establishing a P2P communication path for any two NAT combinations among all NATs including Symmetric NAT is known (for example, Patent Documents). 1).

  In the P2P communication path establishment method described in Patent Document 1, a certain communication terminal predicts a plurality of router port numbers used when a communication terminal that is a connection destination of P2P communication transmits a P2P start request packet. . The certain communication terminal transmits a plurality of packets having the predicted port number as the destination port. Since the router to which the certain communication terminal is connected sets a plurality of port numbers that can receive packets transmitted from the connection destination communication terminal, the possibility that the P2P communication path is successfully established may be improved. it can. Further details will be described below.

  FIG. 25 is a sequence diagram showing a conventional P2P communication path establishment method described in the above-mentioned patent document. In the example shown in FIG. 25, both the router 1 and the router 2 have a Symmetric NAT characteristic.

  First, in steps S501 to S504, as in steps S201 to S204 shown in FIG. 22, each of terminal 1 and terminal 2 transmits an IP / port registration request packet to the server, whereby an IP address and a port number are transmitted. Is registered on the server.

  Next, the terminal 2 transmits a P2P communication request packet to the server to perform P2P communication with the terminal 1 (step S505). In step S505, the terminal 2 changes the source port number from the port number [LP2] used for transmitting the IP / port registration request packet in step S503 to the port number [LP2 + a] (where a is an arbitrary number). Is an integer). The reason why the terminal 2 uses the new port number [LP2 + a] as the source port is that the router 2 assigns a new port number [GP2 + b] to the source (IPL2, LP2 + a) as the source port number is changed. (Where b is an arbitrary integer and an increment value defined in the router 2).

  When the server receives the P2P communication request packet, the server refers to the received packet and transmits the packet including the IP address and the port number (IPG2, GP2 + b) assigned by the router 2 to the source (IPL2, LP2 + a) The notification is transmitted to the terminal 1 (step S506).

  When the terminal 1 receives the IP / port notification packet transmitted from the server, the terminal 1 determines whether or not the terminal 2 is permitted to perform P2P communication, and notifies the server of the determination result (step S507). In step S507, the terminal 1 changes the source port number from [LP1] used for transmitting the IP / port registration request packet in step S501 to [LP1 + c] (where c is an arbitrary integer). ). The reason why the terminal 1 uses the new port number [LP1 + c] as the source port is that the router 1 is expected to assign a new port number [GP1 + d] to the source (IPL, LP1 + c) (however, d is an arbitrary integer and is an increment value defined in the router 1).

  Subsequently, the terminal 1 transmits a P2P start request packet in which (IPG2, GP2 + b + n) is set to the destination based on the IP address and port number (IPG2, GP2 + b) acquired from the received IP / port notification packet to the server. (Step S508). However, n is an arbitrary integer.

  In each of steps S507 and S508, the terminal 1 continuously transmits packets in a very short time. Therefore, when the source port number [LP1 + c] of the packet transmitted in step S507 is converted to [GP1 + d], the source port number [LP1 + c + 1] of the packet transmitted in step S508 is set according to the NAT setting in the router 1. ] Is converted to [GP1 + d + 1].

  Further, as a result of the terminal 1 transmitting the packet in step S508, the router 1 changes the setting so that the response packet can be received. That is, after the router 2 converts the source (IPL1, LP1 + c + 1) of the packet transmitted from the terminal 1 into (IPG1, GP1 + d + 1) and transfers the packet to the router 2 (IPG2, GP2 + b + n), the router 1 Can forward a packet whose source is (IPG2, GP2 + b + n) to the terminal 1 (IPL1, LP1 + c + 1). The reason why the terminal 1 transmits the packet from the acquired destination port number to the port number incremented by an arbitrary integer value n will be described later.

  On the other hand, when the server receives the P2P communication permission packet transmitted from the terminal 1 in step S507, the server transmits a communication permission packet including the IP address and port number (IPG1, GP1 + d) of the router 1 to the terminal 2 (step S509). .

  Next, as soon as the terminal 2 receives the IP address and port number (IPG1, GP1 + d) for accessing the terminal 1, it transmits a P2P start request packet to the terminal 1 (step S510). In step S510, the terminal 2 sends n P2P start request packets having the port number [GP1 + d + 1] of the router 1 as the destination port while incrementing the source port number by 1 from [LP1 + a + 1]. The router 2 converts the source port number of each of the n P2P start request packets transmitted from the terminal 2 to [GP2 + b + m] to [GP2 + m + n−1], respectively. Here, n indicates that the NAT setting can be changed by a packet transmitted from another terminal (not shown) connected to the router 1 during a slight elapsed time during the execution of steps S508 to S510. It is a value determined in consideration of sex. M is an integer of n or less.

  As a result, the destination port number [GP2 + b + n] of the packet transmitted from the terminal 1 in step S508 matches any one of [GP2 + b + m] to [GP2 + b + m + n−1]. Therefore, the router 1 transfers one of the received n P2P start request packets to the terminal 1 (step S510-2). FIG. 25 shows an example in which the port numbers [GP2 + b + n] and [GP2 + b + m + 2] match.

When the terminal 1 receives the P2P start request packet transferred from the router 1, the terminal 1 returns a P2P start response packet (step S511). When the router 2 whose NAT setting has been adjusted in step S510-2 transfers the P2P start response packet transmitted from the terminal 1 to the terminal 2, a P2P communication path is established between the terminal 1 and the terminal 2.
JP 2004-180003 A

  As described above, as a method for establishing a P2P communication path between two communication terminals each connected to different private networks, a method using STUN and a method disclosed in Patent Document 1 are known. Yes.

  However, these conventional methods have the following problems.

  First, in the method using STUN, first, two servers having different IP addresses are required in order to investigate NAT characteristics to which a communication terminal is connected. Therefore, in reality, there is a problem in terms of costs required for server operation and maintenance. Second, the STUN method cannot establish a P2P communication path when either one of two communication terminals is connected to a router having NAT characteristics other than Cone NAT.

  Next, in the technique disclosed in Patent Document 1, the P2P communication path has been successfully established without depending on the NAT characteristics of the router to which the communication terminal is connected (that is, Cone NAT or Symmetric NAT). The possibility to do is improved. More specifically, when the terminal 2 next sends a packet, the terminal 1 determines a port number that may be assigned in the router 1 based on the WAN side port number of the router 2 notified from the server. Predict. The terminal 2 sends out a plurality of packets in order to increase the possibility that the port number assigned by the router 2 matches the port number predicted by the terminal 1.

  However, when both of the two communication terminals are connected to a router having the Full Cone NAT characteristic, it is possible to easily predict the port number. Therefore, it is necessary to transmit a plurality of packets from one communication terminal. There is no. Conversely, when one communication terminal transmits a plurality of packets, the following problem occurs. First, sending a plurality of useless packets increases the time required for establishing a communication path. Second, when a router connected to a communication terminal has an intrusion detection function, when the router receives a plurality of packets destined for a port number not used for communication, the received packet is illegally accessed. There is a possibility of erroneous detection that the packet is an attempt. When a router having an intrusion detection function detects unauthorized access, it discards all received packets, which may cause communication terminals connected to the router to be unable to communicate, resulting in a decrease in connectivity.

  Therefore, an object of the present invention is to provide a communication device, a communication method, and a communication system that can investigate NAT characteristics using a single server having one IP address. It is another object of the present invention to provide a communication apparatus, a communication system, and a communication method capable of establishing a P2P communication path by an optimum connection procedure according to the NAT characteristic based on the investigated NAT characteristic. This improves the connection time and connectivity when connecting two terminals.

  A first aspect of the present invention is directed to a communication device connected to a server via a first relay device having a NAT (network address translation) function. The communication apparatus includes a first transmission process for transmitting at least two address information investigation request packets having the same source port number and different destination port numbers to the server, and the source port number is the first. An address information investigation request transmission unit that executes at least one address information investigation request packet different from the source port number used in the transmission process to the server in an arbitrary order; During the second transmission process, communication for transmitting a communication preparation packet having a transmission source port number identical to any one of the transmission source port numbers used in the first and second transmission processes to a destination different from the server The relay port number converted from the source port address of the address information investigation request packet by the preparation request transmission unit and the NAT function of the first relay device Relay characteristics for determining NAT characteristics of the first relay device based on an address information receiving section that receives at least three address information packets included in the data section from a server and a relay port number included in the received address information packet A determination unit.

  The communication device defined by the first aspect refers to the address information packet returned from the server and acquires the port number set by the first relay device. Therefore, the communication device can determine the NAT characteristic of the first relay device by communicating with one server.

  The second aspect is directed to a communication method for communication by a communication terminal connected to a server via a first relay device having a NAT (network address translation) function. The communication method includes a first transmission process for transmitting at least two address information investigation request packets having the same source port number and different destination port numbers to the server, and the source port number is the first. A second transmission process for transmitting at least one address information investigation request packet different from the transmission source port number used in the transmission process to the server, in any order, and between the first and second transmission processes In addition, a communication preparation packet whose source port number is the same as one of the source port numbers used in the first and second transmission processes is transmitted to a destination different from the server, and the NAT of the first relay device is transmitted. By the function, at least three address information packets including the relay port number converted from the source port address of the address information investigation request packet in the data portion It includes receiving from the server, based on the relay port numbers included in the received address information packet to determine NAT characteristic of the first relay device.

  The communication method defined by the second aspect enables the terminal device to acquire the port number set by the first relay device with reference to the address information packet returned from the server. Therefore, the communication device can determine the NAT characteristic of the first relay device by communicating with one server.

  According to a third aspect, a server, a first relay device having a NAT (network address translation) function, a second relay device having a NAT function, and a first relay device connected to the server via the first relay device. The present invention is directed to a communication system including one communication device and a second communication device connected to a server via a second relay device.

  Each of the first and second communication devices included in the communication system transmits at least two address information investigation request packets having the same source port number and different destination port numbers to the server. And a second transmission process in which at least one address investigation information request packet whose source port number is different from the source port number used in the first transmission process is transmitted to the server in an arbitrary order. Between the address information investigation request transmission unit to be executed and the first and second transmission processes, any one of the transmission source port numbers used in the first and second transmission processes by a destination different from the server A communication preparation request transmission unit that transmits a communication preparation packet that is the same as the port number and the NAT function of the relay device to which each is connected, and an address information investigation request packet. Based on the address information receiving unit that receives from the server at least three address information packets including the relay port number converted from the source port address of the network in the data part, and the relay port number included in the received address information packet And a relay characteristic determining unit that determines the NAT characteristic of the relay apparatus to which each is connected.

  On the other hand, the server included in the communication system includes an address information investigation unit that extracts the relay port number included in the address information investigation request packet transmitted from each of the first and second communication devices, and the extracted relay port number. And an address information transmitting unit that transmits an address information investigation request response packet including the data part in the data part.

  In the communication system defined by the third aspect, each of the first and second communication devices exchanges packets with the server to determine the characteristics of the relay device to which each is connected, and the determination result Is notified to the connection destination communication apparatus via the server. Therefore, each of the first and second communication devices can acquire both the NAT characteristic of the relay device to which the first and second communication devices are connected and the NAT characteristic of the relay device connected to the connection destination communication device.

  According to the communication apparatus, the communication system, and the communication method according to the present invention, it is possible to investigate the NAT characteristic of the router by communicating with a server having one IP address.

  In addition, according to the communication device, the communication system, and the communication method according to the present invention, it is possible to efficiently establish a P2P communication path according to the optimum connection sequence corresponding to the NAT characteristics of the investigated router.

  In addition, according to the communication device, the communication system, and the communication method according to the present invention, when communication terminals connected to different private networks establish a P2P communication path, “reduction of connection time”, “communication impossible” It is possible to achieve “avoidance of state” and “improvement of connectivity”.

  FIG. 1 is a block diagram showing an overall configuration of a communication system according to an embodiment of the present invention.

  The global network 00 and the private network 01 are connected via a router 101 having a global IP address [IPG1]. The global network 00 and the private network 02 are connected via a router 201 having a global IP address [IPG2]. A server 001 having a global IP address [IPS] is connected to the global network 00. A terminal 100 having a local IP address [IPL1] is connected to the private network 01. A terminal 200 having a local IP address [IPL2] is connected to the private network 02.

  FIG. 2A is a block diagram showing a schematic configuration of terminal 100 shown in FIG.

  The terminal 100 includes an address information investigation request transmission unit 1001, an address information reception unit 1002, a direct communication preparation request transmission unit 1003, a relay characteristic determination unit 1004, a relay information transmission unit 1005, a relay information reception unit 1006, A communication control unit 1007 and a communication unit 1008 are provided.

  The address information investigation request transmission unit 1001 transmits a port number investigation request packet to the server 001 in order to investigate the port number set in the packet transmitted from the terminal 100 by the NAT function of the router 101. More specifically, the address information investigation request transmission unit 1001 includes the first transmission process for transmitting at least two packets having different destination port numbers and the one in which the source port number is used in the first transmission process. And a second transmission process for transmitting at least one different packet. In the first transmission processing, it is more preferable that the address information investigation request transmission unit 1001 transmits two packets having the same source port and different destination port numbers. Further, the execution order of the first and second transmission processes may be arbitrary.

  The address information receiving unit 1002 receives from the server 001 a port number investigation response packet returned from the server 001 in response to the port number investigation request packet. The port number investigation response packet includes the port number set by the NAT function of the router 101 in the data portion.

  The direct communication preparation request transmitting unit 1003 prepares a port to be used for P2P communication on the router 101 between the time when one of the first and second transmission processes is executed and the time when the other is executed. A P2P communication preparation packet is transmitted to other terminals connected to the global network 00.

  The relay characteristic determining unit 1004 determines the NAT characteristic of the router 101 based on the port number included in the data part of the port number investigation response packet received by the address information receiving unit 1002.

  The relay information transmission unit 1005 transmits a NAT information notification packet to the server in order to notify the NAT characteristic of the router 101 determined by the relay characteristic determination unit 1004 to the connection destination terminal 200 via the server.

  The relay information receiving unit 1006 receives the NAT information notification packet transmitted by the connection destination device. The received NAT information notification packet includes information necessary for specifying the NAT characteristic of the router 201 connected to the terminal 200 of the connection destination.

  Based on the combination of the NAT characteristics of the router 101 to which the terminal 100 is connected and the NAT characteristics of the router 201 to which the communication destination terminal 200 is connected, the communication control unit 1007 is based on a plurality of sequences prepared in advance. , Select and execute an optimal sequence for establishing P2P communication.

  The communication unit 1008 transmits the packet generated in each of the above units to the private network 01 and receives the packet addressed to the terminal 100 via the private network 01.

  FIG. 2B is a block diagram showing a schematic configuration of terminal 200 shown in FIG.

  The terminal 200 includes an address information investigation request transmission unit 2001, an address information reception unit 2002, a direct communication preparation request transmission unit 2003, a relay characteristic determination unit 2004, a relay information transmission unit 2005, a relay information reception unit 2006, A communication control unit 2007 and a communication unit 2008 are provided. Since each function of address information investigation request transmission unit 2001 to communication unit 2008 is the same as each function of address information investigation request transmission unit 1001 to communication unit 1008 shown in FIG. 2A, description thereof will not be repeated here. .

  FIG. 2C is a block diagram showing a schematic configuration of the server 001 shown in FIG.

  The server 001 includes an address information investigation unit 0011, an address information transmission unit 0012, a relay information transfer unit 0013, and a communication unit 0014.

  When receiving the port number investigation request packet from each of the terminals 100 and 200, the address information investigation unit 0011 extracts the transmission source port number included in the header of the received packet. The source port number extracted by the address information examining unit 0011 is a value set by the NAT function of the router 101 or 201. The address information examining unit 0011 outputs the extracted port number to the address information transmitting unit 0012.

  The address information transmission unit 0012 generates a port number investigation response packet including the port number received from the address information investigation unit 0011 in the data part, and returns the generated packet to the terminal.

  When the relay information transfer unit 0013 receives the NAT information notification packet from one of the two terminals, the relay information transfer unit 0013 transfers the received NAT information notification packet to the other terminal.

  The communication unit 0014 transmits the packet generated in each unit to the global network 00, and receives a packet addressed to the server 001 via the global network 00.

  Here, a control sequence in the communication system according to the present embodiment will be described.

  FIG. 3 is a sequence diagram illustrating an outline of processing executed by the communication system according to the present embodiment to establish a P2P communication path.

  As shown in FIG. 3, the P2P communication path establishment method according to the present embodiment is composed of four phases, that is, a NAT characteristic investigation phase S801, a preparation phase S802, an information exchange phase S803, and a P2P communication path establishment phase. .

  In the NAT characteristic investigation phase S801, the terminal 100 communicates with the server 001 and investigates the NAT characteristic (for example, Cone NAT or Symmetric NAT) of the router 101 connected to the terminal 100. Further, the terminal 100 investigates whether or not the router 101 has the Port Reuse characteristic. Similarly, the terminal 200 communicates with the server 001 and investigates the NAT characteristic of the router 101 and the presence / absence of the Port Reuse characteristic.

  In the preparation phase S802, the terminal 100 transmits a packet to the router 101 to which the terminal 100 is connected, and opens a port of the router 101 used for communication with the terminal 200. Similarly, the terminal 200 transmits a packet to the router 201 and opens a port of the router 201 used for communication with the terminal 100.

  In the information exchange phase S803, each of the terminals 100 and 200 exchanges the NAT characteristic determined in the NAT characteristic investigation phase and information necessary for establishing the P2P communication path via the server 001.

  In the P2P communication path establishment phase S804, each of the terminals 100 and 200 transmits a packet to each other based on the information acquired in the information exchange phase S803 to establish a P2P communication path.

  In FIG. 3, the order of the NAT characteristic investigation phase S801 and the preparation phase S802 is specified for convenience of illustration, but the NAT characteristic investigation phase S801 and the preparation phase S802 are completed before the information exchange phase S803. It should be.

  Hereinafter, the four phases will be described in more detail.

<NAT characteristics investigation phase and preparation phase>
FIG. 4A is a sequence diagram showing details of processing in the NAT characteristic investigation phase and the preparation phase shown in FIG.

  First, the terminal 100 transmits a port number investigation request 1 to the server 001 (step S901). More specifically, the address information investigation request transmission unit 1001 has a port in which the port number [LP1] of the terminal 100 is set as the source port number and the port number [SP0] of the server 001 is set as the destination port number. Number search request packet 1 is transmitted.

  Next, the server 001 returns the port number investigation response 1 to the terminal in response to the port number investigation request 1 transmitted from the terminal 100 (step S902). More specifically, the communication unit 0014 receives the port number investigation request packet 1 and transfers the received packet to the address information investigation unit 0011. The address information examining unit 0011 refers to the header of the transferred packet, extracts the source port number, that is, the port number [GP11] set by the router 101, and addresses the extracted source port number [GP11]. The information is output to the information transmission unit 0012.

  The address information transmission unit 0012 generates a port number investigation response packet 1 including the port number [GP11] of the router 101 received from the address information investigation unit, and outputs the generated packet to the communication unit 0014. The communication unit 0014 returns the packet received from the address information transmission unit 0012 to the terminal 100. In terminal 100, upon receiving port number investigation response packet 1, communication unit 1008 transfers the received packet to address information receiving unit 1002.

  Next, the terminal 100 transmits the port number investigation request 2 to the server 001 (step S903). More specifically, the address information investigation request transmission unit 1001 has a port in which the port number [LP1] of the terminal 100 is set as the source port number and the port number [SP1] of the server 001 is set as the destination port number. Number search request packet 2 is transmitted.

  Next, the server 001 returns the port number investigation response 2 to the terminal in response to the port number investigation request 2 transmitted from the terminal 100 (step S904). More specifically, the communication unit 0014 receives the port number investigation request packet 2 and transfers the received packet to the address information investigation unit 0011. The address information investigation unit 0011 refers to the header of the transferred packet, extracts the source port number, that is, the port number [GP12] of the router 101, and uses the extracted source port number [GP12] as the address information transmission unit. To 0012.

  Next, the terminal 100 transmits a P2P communication preparation packet to the terminal 200 (step S905). More specifically, the terminal 100 transmits a packet destined for the IP address and port number of the terminal 200 to the router 101 in order to prepare a port number [GP13] used when performing P2P communication with the terminal 200. The terminal 100 may acquire the IP address and port number of the terminal 200 in advance. Also, the server 001 sets the IP address and port number of the terminal 200 in the port number investigation response packet 1 or 2, and the terminal 100 refers to the received port number investigation response packet to refer to the IP address and port number of the terminal 200. You may get

  Here, the destination port number of the packet transmitted by the terminal 100 is the port number actually used by the router 201 (for example, when the terminal 200 is communicating with the server 001, the router 201 transmits the packet of the terminal 200). (Assigned port number). This is because if the router 201 has an intrusion detection function and receives a P2P communication preparation packet in which a port number not used for communication by the router 201 is set as a destination port number from the terminal 100, the packet is transmitted as an unauthorized access packet. May be falsely detected. When the router 201 detects an unauthorized access packet, it discards all subsequent packets transmitted from the terminal 100, so that the terminal 100 may be unable to communicate with the terminal 200.

  Alternatively, the terminal 100 may adjust the TTL (Time To Live) value of the packet so that the P2P communication preparation packet does not reach the router 201. The terminal 100 transmits the P2P communication preparation packet in order to prepare a port number for P2P communication with the terminal 200 in the router 101 to which the terminal 100 is connected, and the P2P communication preparation packet necessarily reaches the router 201. There is no need. Therefore, the terminal 100 may adjust the TTL value so that the P2P communication preparation packet reaches the router 101 and does not reach the router 201.

  In step S905, the router 101 sets NAT so that the packet transmitted from the terminal 200 to the router 101 (IPG1, GP13) via the router 201 can be accepted. The method of setting the NAT by the router 101 differs depending on the NAT characteristics (FIGS. 18A to 18E), but the description here is omitted.

  Next, the terminal 100 transmits the port number investigation request 3 to the server 001 (step S906). More specifically, the address information investigation request transmission unit 1001 has a packet in which the port number [LP2] of the terminal 100 is set as the source port number and the port number [SP0] of the server 001 is set as the destination port number. Is transmitted as a port number investigation request 3.

  In response to the port number investigation request 3 transmitted from the terminal 100, the server 001 returns the port number investigation response 3 to the terminal 100 (step S907). More specifically, the communication unit 0014 receives the port number investigation request packet 2 and transfers the received packet to the address information investigation unit 0011. The address information investigation unit 0011 refers to the header of the transferred packet, extracts the source port number, that is, the port number [GP14] of the router 101, and uses the extracted source port number [GP14] as the address information transmission unit. To 0012.

  Upon receiving the port number investigation response packets 1 to 3, the terminal 100 determines the NAT characteristic of the router 101 (step S908). More specifically, in the terminal 100, the relay characteristic determining unit 1004 performs the following based on the three port numbers [GP11], [GP12], and [GP14] of the router 101 included in the port number investigation response packets 1 to 3. The NAT characteristic of the router 101 is determined in accordance with the processing flow.

  FIG. 4B is a flowchart showing details of the process in step S908 shown in FIG. 4A.

  The relay characteristic determination unit 1004 first determines whether or not the acquired port numbers [GP11] and [GP12] match (step S909). If the port number [GP11] matches [GP12] (Yes in step S909), the relay characteristic determining unit 1004 proceeds to step S911. On the other hand, when the port number [GP11] does not match [GP12] (No in step S909), the relay characteristic determining unit 1004 determines that the NAT of the router 101 is PS NAT (step S910), and proceeds to step S914. .

  When [GP11] matches [GP12] (Yes in step S909), the relay characteristic determination unit 1004 calculates the absolute value β of the difference between the acquired port numbers [GP12] and [GP14], and β is 1. It is determined whether or not (step S911). If β is 1, the relay characteristic determining unit 1004 determines that the NAT of the router 101 is Cone NAT (step S912), and proceeds to step S914. In other cases, the relay characteristic determining unit 1004 determines that the NAT of the router 101 has either the Cone or AS NAT characteristic (step S913), and proceeds to step S914.

  Further, the relay characteristic determining unit 1004 determines whether or not the port number [GP11] of the router 101 acquired in step S902 matches the source port number [LP1] of the port number investigation request packet 1 transmitted in step S901. Judgment is made (step S914). When [GP11] matches [LP1] (step S914), the relay characteristic determining unit 1004 determines that the router 101 has the Port Reuse characteristic (step S915), and ends the NAT characteristic determining process. In other cases, the relay characteristic determining unit 1004 ends the NAT characteristic determining process.

  Note that, in order to strictly determine whether or not the router 101 has Port Reuse, the relay characteristic determination unit 1004 may further determine whether or not the port number [GP14] matches [LP2]. . In this case, the relay characteristic determining unit 1004 determines that the router 101 has the Port Reuse characteristic when [GP11] matches [LP1] and [GP14] matches [LP2]. . Further, in order to gently determine whether or not the router 101 has the Port Reuse characteristic, either [GP11] matches [LP1] or [GP14] matches [LP2]. In this case, the relay characteristic determining unit 1004 may determine that the router 101 has the Port Reuse characteristic. Further, the processing for determining the Port Reuse characteristic in steps S914 to S915 may be performed before step S909.

  Through the above NAT characteristic determination process, the terminal 100 can investigate the NAT characteristic of the router 101.

  Here, a specific example of the NAT function characteristic determination process will be described.

Figure 5A is a view for explaining the step S910 shown in FIG. 4 B.

  It is determined that the relay characteristic determination unit 1004 determines No in step S909, that is, “GP11” is different from [GP12]. The router 101 corresponds to each of the destination port numbers [SP0] and [SP1]. Therefore, the relay characteristic determining unit 1004 can determine that the router 101 has the PS NAT characteristic.

Figure 5B is a view for explaining the step S912 shown in FIG. 4 B.

  When the relay characteristic determining unit 1004 determines Yes in step S911, that is, determining that the absolute value β of the difference between [GP12] and [GP14] is 1, the destination port number of the P2P communication preparation packet Indicates that the port number [GP13] assigned by the router matches [GP11] and [GP12]. Therefore, the relay characteristic determining unit 1004 can determine that the router 101 has the Cone NAT characteristic.

Each of FIGS. 5C and 5D are diagrams for explaining the step S913 shown in FIG. 4 B.

  If the relay characteristic determining unit 1004 determines No in step S911, that is, if the absolute value β of the difference between [GP12] and [GP14] is not 1, the router is set as the destination port number of the P2P communication preparation packet. The port number [GP13] assigned by is coincident with [GP11] and [GP12] (FIG. 5C) and may not coincide (FIG. 5D). Therefore, the relay characteristic determining unit 1004 determines that the router 101 has either the Cone or AS NAT characteristic.

  The terminal 200 investigates the NAT characteristics of the router 201 according to the same procedure as that shown in FIGS. 4A and 4B, and the router 201 opens a port for use in P2P communication.

  4A also shows the NAT characteristic investigation phase (steps S901 to S904, steps S906 to S908) and the preparation phase (step S902) for convenience of explanation, the NAT characteristic investigation phase, the preparation phase, May be separated from each other.

  Further, in step S913, when the relay characteristic determining unit 1004 is not β, Cone NAT and AS NAT are not distinguished from each other in the P2P communication path establishment phase described later without distinguishing between them. Because it is possible to do. However, the terminal 100 may further execute the following process in order to distinguish between Cone NAT and AS NAT.

  First, in the terminal 100, the address information investigation request transmission unit 1001 sets an IP address different from the IP address [IPS] of the server 001 as the destination address and sets the port number [LP1] as the source address. The packet is further transmitted as a port number check request. The address information receiving unit receives a response packet to the port number investigation request packet, and determines whether or not the port number [GP15] assigned to the port number investigation request packet by the router 101 matches [GP11]. The relay characteristic determining unit 1004 can determine that the router 101 has the Cone NAT characteristic when [GP15] matches [GP11], and otherwise, the router 101 has the AS NAT characteristic. Judgment can be made.

  Further, the NAT characteristic investigation phase and the preparation phase are not necessarily executed in the order shown in the present embodiment. Therefore, the terminals 100 and 200 may execute the steps in a different order from the present embodiment as long as the NAT characteristic investigation phase and the preparation phase are completed before the subsequent information exchange phase.

<Information exchange phase>
FIG. 6 is a sequence diagram showing details of processing in the information exchange phase shown in FIG.

  First, the terminal 200 notifies the server 001 of NAT information indicating the NAT characteristics of the router 201 to which the terminal 200 is connected (step S1001). More specifically, the relay information transmitting unit 2005 generates a packet including NAT information indicating the NAT characteristic of the router 201 acquired in the NAT characteristic investigation phase in the data part, and the generated packet is transmitted to the server 001 via the communication unit 2008. Send to.

  The NAT information includes information indicating (a) Cone NAT characteristics, (b) Cone or AS NAT characteristics, and (c) PS NAT characteristics, and (d) information indicating presence / absence of Port Reuse characteristics. It is out.

The NAT information further includes parameters shown below according to the NAT characteristics.
(A) When the NAT characteristic is Cone, the NAT information includes at least the value of the port number [GP22] (= [GP24]).
(B) When the NAT characteristic is Cone or AS, the NAT information includes at least the values of the port numbers [GP22] and [GP24].
(C) When the NAT characteristic is PS, the NAT information includes at least a difference α between the port numbers [GP21] and [GP22] and a port number [GP24].
(D) In addition to the above (a) to (c), when the NAT has a Port Reuse characteristic, the terminal 200 includes at least a port number [LP4] that the terminal 200 plans to newly open in the next P2P communication path establishment phase . When the router 201 has the Symmetric NAT characteristic, the port number [LP4] needs to be a new number, but when the router 201 has the Cone NAT characteristic, the port number [LP4] is [GP21]. To [GP24].

  When the server 001 receives the NAT information notification packet, the server 001 transfers the received NAT information notification packet to the terminal 100 (step S1002). More specifically, in the server 001, the relay information transfer unit 0013 receives the NAT information notification packet transmitted from the terminal 200 through the communication unit 0014, and receives the received NAT information notification packet through the communication unit 0014. To 100.

  Next, the terminal 100 notifies the server 001 of NAT information indicating the NAT characteristics of the router 101 to which the terminal 100 is connected (step S1003). More specifically, the relay information transmission unit 1005 generates a packet including NAT information indicating the NAT characteristic of the router 101 acquired in the NAT characteristic investigation phase, and transmits the generated packet to the server 001 via the communication unit 1008. . Since the NAT information transmitted from the terminal 100 to the server 001 is the same as that transmitted from the terminal 200 to the server 001, description thereof will not be repeated here.

  When the server 001 receives the NAT information notification packet, the server 001 transfers the received NAT information notification packet to the terminal 200 (step S1004). More specifically, in server 001, relay information transfer unit 0013 receives a NAT information notification packet transmitted from terminal 100 via communication unit 0014, and receives the received NAT information notification packet via communication unit 0014. 200.

  In this embodiment, each of the terminals 100 and 200 notifies the server 001 of NAT characteristics and parameters corresponding to the NAT characteristics as NAT information, but the three port numbers (port number [GP11 ], [GP12] and [GP14] or the port numbers [GP21], [GP22] and [GP24]) may be notified to the server 001. In this case, each of the terminals 100 and 200 determines the NAT characteristics of each of the routers 101 and 201 based on the three port numbers acquired from the server 001 according to the processing flow shown in FIG. Calculate the value. Alternatively, the server 001 determines the NAT characteristics of each of the routers 101 and 201 based on the three port numbers acquired from the terminals 100 and 200, respectively, according to FIG. 4B. The characteristics may be notified to the terminals 200 and 100, respectively.

  In the above description, the NAT information includes information exemplified in items (a) to (d), but may include other information. For example, the NAT information may include two values defined by Δ1 = | [GP12] − [GP11] | and Δ2 = | [GP14] − [GP12] |. In this case, the relay characteristic determination unit acquires the values of Δ1 and Δ2, and determines that the router has the Cone NAT characteristic when Δ1 = 0 and Δ2 = 0 are satisfied, and Δ2> 1 is satisfied. In addition, it may be determined that the router has the AS NAT characteristic, and if Δ1 ≠ 0 is satisfied, it may be determined that the router has the PS NAT characteristic.

  The NAT information may include information other than those described above as long as at least one of the terminal 100, the terminal 200, and the server 001 is useful information for determining the NAT characteristics of the router.

  Further, in the present embodiment, the terminal 200 transmits the NAT information notification packet first, but the terminal 100 may transmit the NAT information notification packet first.

<P2P communication channel establishment phase>
FIG. 7 is a flowchart showing processing executed by the terminal 100 in the P2P communication path establishment phase shown in FIG. In the P2P communication path establishment phase, the terminals 100 and 200 attempt to establish a P2P communication path using the parameters acquired in the information exchange phase.

  First, the communication control unit 1007 determines whether or not the router 201 has a Cone NAT characteristic (step S1201). When the communication control unit 1007 determines that the router 201 has the Cone NAT characteristic (Yes in step S1201), the communication control unit 1007 proceeds to step S1207 described later, and the port number [GP22] of the router 201 is set as the destination port number. A P2P start request packet is transmitted.

  If the router 201 does not have the Cone NAT characteristic (No in step S1201), the communication control unit 1007 determines whether the router 201 has both the Cone or AS NAT characteristics (step S1202). If the communication control unit 1007 determines that the router 201 may be either Cone or AS NAT (Yes in step S1202), the communication control unit 1007 proceeds to step S1208, which will be described later, and is greater than or equal to [GP22] and less than [GP24] A plurality of P2P communication start request packets having a range value set as the destination port number are transmitted.

  When the router 201 does not have either the Cone or AS NAT characteristics (No in step S1202), the communication control unit 1007 determines whether or not the router 201 has the Port Reuse characteristics (step S1203). When the router 201 has the Port Reuse characteristic (Yes in step S1203), the communication control unit 1007 does not consider the NAT characteristic of the router 201 and the P2P start request in which the port number [LP4] is set as the destination port number. Send the packet. If the router 201 does not have the Port Reuse characteristic (No in step S1203), the communication control unit 1007 proceeds to step S1204.

  The communication control unit 1007 reaches step S1204 when the router 201 has neither the Cone NAT characteristic nor the AS NAT characteristic, that is, when the router 201 has the PS NAT characteristic. When the router 201 is PS NAT, the terminal 100 needs to select a process in consideration of the characteristics of the router 101 to which the terminal 100 is connected as shown below.

  The communication control unit 1007 determines whether or not the router 101 is identified as having the Cone NAT characteristic (step S1204). If the router 101 has the Cone NAT characteristic (Yes in step S1204), the communication control unit 1007 proceeds to step S1210 described later, and transmits a P2P start request packet in which the port number [GP24 + α2] is set as the destination port number.

  If it is not specified that the router 101 has the Cone NAT characteristic (No in step S1204), the communication control unit 1007 determines whether the router 101 has either the Cone or AS NAT characteristic (step S1205). . If the router 101 has either Cone or AS NAT characteristics (Yes in step S1205), the communication control unit 1007 proceeds to step S1211 described later, and is greater than or equal to [GP24 + α2] and less than [GP24 + (n−1) × α2]. A plurality of P2P start request packets in which the value in the range is set as the destination port number is transmitted (where n is an arbitrary positive integer).

  When the router 101 does not have either the Cone NAT characteristic or the AS NAT characteristic (No in step S1205), the communication control unit 1007 determines whether the router 101 has the PS NAT characteristic (step S1206). ). If the router 101 has PS NAT characteristics (Yes in step S1206), the communication control unit 1007 proceeds to step S1212 described later, and transmits a P2P communication start packet in which the port number [GP24 + α2] is set as the destination port number. Send. If the router 101 does not have PS NAT characteristics, the process returns.

  In the P2P communication path establishment phase, the process executed by the communication control unit 2007 of the terminal 200 is the same as the process executed by the terminal 200, and thus detailed description thereof is omitted.

  Further, in the sequence diagram shown in FIG. 7, the order of each determination process is specified as an example, but each determination process may be executed in an order other than that shown in FIG. 7. For example, it may be determined first whether the router has the Port Reuse characteristic (step S1203).

  Hereinafter, details of each of steps S1207 to S1212 described above will be described.

<Connection Sequence 1: When Router 201 is Cone NAT>
FIG. 8 is a diagram showing details of the connection sequence 1 shown in FIG. In FIG. 8, it is assumed that the router 101 has a Cone NAT characteristic for the sake of simplicity.

  As described above, the terminal 100 determines that the router 201 to which the terminal 200 is connected has the Cone NAT characteristic based on the information notified from the terminal 200 in steps S1001 and S1002 of the information exchange phase (FIG. 6), and NAT. In the characteristic investigation phase and the preparation phase, the port numbers [GP22] and [GP24] opened by the router 201 are grasped. Since the router 201 has the Cone NAT characteristic, the terminal 100 can predict that the port number of the router 201 that the terminal 200 uses for P2P communication is [GP22].

  Therefore, the communication control unit 1007 transmits, as a P2P start request, a packet in which [LP1] is set as the source port number and [GP22] is set as the destination port number (step S1301). The router 101 assigns a port number [GP13] to the packet transmitted from the terminal 100.

  At this stage, since the router 201 does not set NAT so as to transfer the packet in which the destination port number transmitted from the terminal 100 is set to [GP22], the router 201 transmits the P2P transmitted in step S1301. The start request packet is discarded and not transferred to the terminal 200. At this stage, the router 101 sends a packet in which the port number [GP22] of the router 201 is set as the source port number and (IPG1, GP13) is set as the destination IP address and the destination port number to the terminal 100 ( Set NAT to forward to IPL1, LP1).

  On the other hand, based on the information notified from the terminal 100 in the information exchange phase, the terminal 200 includes information indicating that the router 101 to which the terminal 100 is connected has the Cone NAT characteristic, and the router 200 in the NAT characteristic investigation phase and the preparation phase. Information indicating the port numbers [GP12] and [GP14] opened by 101 is acquired. When the terminal 200 acquires information on the NAT characteristic of the router 101, the terminal 200 predicts that the port number used for the P2P communication by the router 101 is [GP12] based on the fact that the router 101 has the Cone NAT characteristic. Then, a packet in which [LP1] is set as the source port number and [GP12] is set as the destination port number is transmitted as a P2P start request (step S1302). Since the router 101 has a Cone NAT characteristic, the port number [GP12] is the same as [GP13].

  The router 101 transfers the P2P start request packet transmitted from the terminal 200 to the terminal 100. In terminal 100, communication control unit 1007 transmits a P2P start response packet to terminal 200 via communication unit 1008 in response to the P2P start request packet transferred from router 101 (step S1303).

  When the router 201 has the Cone NAT characteristic, a P2P communication path can be established between the terminals 100 and 200 according to the above processing procedure.

<Connection Sequence 2: When Router 201 has either Cone or AS NAT characteristics>
9A is a diagram showing details of the connection sequence 2 shown in FIG. 7, and FIG. 9B is a diagram showing the port numbers shown in FIG. 9A. In FIG. 9A, for simplification of description, it is assumed that the router 101 has a Cone NAT characteristic.

  First, when the situation shown in FIG. 5A is applied to the terminal 200, when the router 201 has the Cone NAT characteristic, it is predicted that the port number assigned to the P2P communication by the router 201 is [GP22].

  Next, when the situation shown in FIG. 5D is applied to the terminal 200, when the router 201 has AS NAT characteristics, the port number assigned to the P2P communication by the router 201 is between [GP22] and [GP24]. Any one port number.

  Therefore, in the terminal 100, the communication control unit 1007 has a plurality of transmission port numbers each set with [LP1] and each destination port number with a port number greater than or equal to [GP22] and less than [GP24]. A P2P start request packet is transmitted (step S1401). The router 101 assigns the port number [GP13] prepared in the preparation phase to all P2P start request packets.

  At the time when the P2P start request packet is transmitted in step S1401, the router 201 sets NAT so as to receive a packet in which each port number of [GP22] to [GP24-1] is set as the destination port number. Not done. Therefore, at this stage, the router 201 discards each received P2P start request packet and does not transfer it to the terminal 200. It should be noted that the router 101 transmits each packet in which (IPG1, GP13) is set as the destination and each port number in the range of [GP22] to [GP24-1] is set as the source port number to the terminal 100. Set NAT to forward to (IPL1, LP1).

  On the other hand, based on the information notified from the terminal 100 in the information exchange phase, the terminal 200 includes information indicating that the router 101 to which the terminal 100 is connected has the Cone NAT characteristic, and the router 200 in the NAT characteristic investigation phase and the preparation phase. 101 indicates information indicating the port numbers [GP12] and [GP14] opened by 101.

  When the terminal 200 acquires information on the NAT characteristic of the router 101, the terminal 200 predicts that the port number used for the P2P communication by the router 101 is [GP12] based on the fact that the router 101 has the Cone NAT characteristic. Then, a packet in which [LP1] is set as the source port number and [GP12] is set as the destination port number is transmitted as a P2P start request (step S1402). Since the router 101 has a Cone NAT characteristic, the port number [GP12] is the same as [GP13].

  In step S1401, the router 101 has already set NAT so as to be able to accept a packet whose source port number is [GP22 + n] by transmitting a P2P start request packet (where [ GP22] ≦ [GP22 + n] <[GP24] and [GP22 + n] = [GP23] is established).

  Therefore, the router 101 transfers the P2P start request packet transmitted from the terminal 200 to the terminal 100. In terminal 100, communication control unit 1007 transmits a P2P start response packet to terminal 200 through communication unit 1008 in response to the P2P start request packet transferred from router 101 (step S1403).

  When the router 201 does not specify either Cone or AS NAT characteristics, a P2P communication path can be established between the terminals 100 and 200 according to the above processing procedure.

  As a modification, the following processing may be executed instead of step S1401 described above.

  FIG. 10 is a diagram for explaining a port number allocation interval of a router.

  In the example shown in FIG. 9A above, the terminal 100 transmits the P2P start request packet while incrementing the destination port number by 1 within the range of [GP22] or more and less than [GP24]. For this reason, the increment of the destination port number may not be 1.

  For example, a certain type of router has a function of detecting unauthorized intrusion by determining whether a destination port number of a received packet conforms to a predetermined port number allocation rule. If the router 201 has this type of intrusion detection function, the intrusion detection function may be activated when the router 201 receives a P2P start request packet transmitted from the terminal 100 while incrementing the destination port number by one. is there. It is conceivable that the terminal 100 cannot communicate with the terminal 200 due to the operation of the intrusion detection function.

  Therefore, in consideration of the case where the router 201 has an intrusion detection function, it is conceivable that the terminal 200 obtains a difference in port numbers assigned to the router 201 in advance.

  More specifically, as shown in FIG. 10, when it is determined that the router 201 has either the Cone or AS NAT characteristic in the NAT characteristic investigation phase, the address information investigation request transmission unit 2001 displays the destination port number. Two or more port number investigation request packets in which a certain port number (for example, [SP0]) of the server 001 is set are transmitted while incrementing the source port number by one.

  In response to each received port number investigation request packet, server 001 transmits a response packet including the source port number converted by router 201 to terminal 200. In terminal 200, relay information transmitting section 2005 refers to each response packet transmitted from server 001, and checks the increment of the port number assigned to each of the port number investigation response packets by router 201. The terminal 200 transmits the minimum increment of the port number as a port allocation interval (Δ) to the terminal 100 using a NAT information notification packet.

  In step S1401, the communication control unit 1007 of the terminal 100 may transmit a plurality of P2P start request packets while incrementing the port allocation interval Δ by the destination port number in the range of [GP22] to less than [GP24].

<Connection Sequence 3: When Router 201 has Port Reuse Characteristics>
FIG. 11 is a diagram showing details of the connection sequence 3 shown in FIG. In FIG. 11, for the sake of simplicity, it is assumed that the router 101 also has a Port Reuse characteristic.

  As described above, the terminal 100 determines that the router 201 to which the terminal 200 is connected has the Port Reuse characteristic based on the information notified from the terminal 200 in steps S1001 and S1002 of the information exchange phase (FIG. 6), and NAT. In the characteristic investigation phase and the preparation phase, the router 201 acquires information indicating the port number [LP4] that is scheduled to be opened for P2P communication. Since the router 201 has a Port Reuse characteristic, when the terminal 100 transmits a packet in which [LP4] is set as the source port number, the router 201 assigns the port number [LP4] for the transfer of the packet. Is expected.

  Therefore, the communication control unit 1007 transmits a P2P start request packet in which the port number [LP3] notified to the terminal 200 in steps S1003 and S1004 is set as the source port number and [LP4] is set as the destination port number. (Step S1501). Since the router 101 has the Port Reuse characteristic, the router 101 assigns the same port number as the transmission port number [LP3] of the P2P start request packet to the P2P start request packet.

  At the time when the P2P start request packet is transmitted in step S1501, the router 201 does not set NAT so that the packet in which [LP4] is set as the destination port number can be transferred to the terminal 200. Therefore, the router 201 discards the P2P start request packet and does not transfer the packet to the terminal 200. On the other hand, the router 101 forwards the P2P start request packet to transfer a packet in which (IPG1, LP3) is set as the destination and [LP4] is set as the destination port number to the terminal 100 (IPL1, LP3). Set NAT to forward to.

  On the other hand, the terminal 200 indicates information indicating that the router 101 to which the terminal 100 is connected has the Port Reuse characteristic and the port number [LP3] of the terminal 100 based on the information notified from the terminal 100 in the information exchange phase. Get information and.

  As soon as the terminal 200 acquires information on the NAT characteristics of the router 101, the communication control unit 2007 transmits a P2P start request packet in which [LP3] is set as the destination port number (step S1502).

  The router 101 transfers the P2P start request packet transmitted from the terminal 200 to the terminal 100. In terminal 100, communication control unit 1007 transmits a P2P start response packet to terminal 200 through communication unit 1008 in response to the P2P start request packet transferred from router 101 (step S1503).

  When the router 201 has the Port Reuse characteristic, a P2P communication path can be established between the terminals 100 and 200 according to the above processing procedure.

  In the sequence diagram illustrated in FIG. 11, the terminal 100 transmits the P2P start request before the terminal 200, but the terminal 200 may transmit the P2P start request before the terminal 100.

  Further, in the information exchange phase, the terminals 100 and 200 may notify each other of a plurality of port numbers in order to improve the possibility of successfully establishing the P2P communication path in the P2P communication path establishment phase.

<Connection Sequence 4: When Router 201 has PS NAT characteristics and Router 101 has Cone NAT characteristics>
12A is a diagram showing details of the connection sequence 4 shown in FIG. 7, and FIG. 12B is a diagram showing the port numbers shown in FIG. 12A.

  As described above, the terminal 100 determines that the router 201 to which the terminal 200 is connected has the PS NAT characteristic and the port number [GP24] opened by the router 201 in steps S1001 and S1002 of the information exchange phase (FIG. 6). And NAT characteristic information including the value of the port allocation interval α of the router 201 is acquired. In PS NAT, every time a new port is opened, the port number increases by α, so the terminal 100 can predict that the port used by the router 201 for P2P communication is [GP24 + α].

  Therefore, the communication control unit 1007 of the terminal 100 transmits a P2P start request packet in which [GP24 + α] is set as the destination port number and [LP1] is set as the source port number (step S1601). Since the router 101 has the Cone NAT characteristic, the port number [GP13] prepared in the preparation phase is assigned to the P2P communication start packet.

  At the time when the P2P start request packet is transmitted in step S1601, the router 201 has not set NAT so as to transfer the packet in which [GP24 + α] is set to the destination port number to the terminal 200. The received packet is discarded and not transferred to the terminal 200. On the other hand, the router 101 sets NAT so as to forward a packet in which (IPG1, GP13) is set as the destination and [GP24 + α] is set as the source port number to the terminal 100 (IPL1, LP1). .

  On the other hand, in the information exchange phase, the terminal 200 receives NAT characteristic information including that the router 101 to which the terminal 100 is connected has the Cone NAT characteristic and the port numbers [GP12] and [GP14] opened by the router 101. Have acquired.

  When the terminal 200 acquires information on the NAT characteristic of the router 101, the terminal 200 predicts that the port number used for the P2P communication by the router 101 is [GP12] based on the fact that the router 101 has the Cone NAT characteristic. Then, a P2P start request packet in which [LP1] is set as the source port number and [GP12] is set as the destination port number is transmitted (step S1602). Since the router 101 has a Cone NAT characteristic, the port number [GP12] is the same as [GP13].

The router 101 transfers the P2P start request packet transmitted from the terminal 200 to the terminal 100. In the terminal 100, the communication control unit 1007 transmits a P2P start response packet to the terminal 200 via the communication unit 1008 in response to the P2P start request packet transferred from the router 101 (step S1603).
When the router 201 has the PS NAT characteristic and the router 101 has the Cone NAT characteristic, a P2P communication path can be established between the terminals 100 and 200 according to the above processing procedure.

  In this embodiment, the terminal 100 transmits one P2P start request packet whose destination port number is [GP24 + α]. However, the terminal 100 may transmit a plurality of P2P start request packets. For example, in the NAT information notification phase, while the terminal 100 is exchanging information with the terminal 200, another application of the terminal 200 may occupy the port number [GP24 + α] of the router 201 first. In consideration of this possibility, in order to improve the certainty of establishing the P2P communication path, in addition to [GP24 + α], a plurality of destination port numbers incremented by α, such as [GP24 + 2α] and [GP24 + 3α] The P2P start request packet may be transmitted.

  Further, the following modification may be applied to the processing in step S1601.

  FIG. 12C is a diagram illustrating a relationship of port numbers between terminals and routers.

  As illustrated in FIG. 12C, the terminal 100 transmits a P2P communication start request in which [GP24 + nα] (where n is an integer) that is intentionally largely shifted from [GP24] is set as the transmission port number. . For example, n is set to 5. After that, the terminal 200 transmits n request packets in which [GP12] is set as the destination port number. Of the n request packets transmitted from the terminal 200, one packet whose source port number is converted to [GP24 + nα] is transferred to the terminal 100 by the router 101, thereby establishing a P2P communication path.

<Connection Sequence 5: When Router 201 has PS NAT characteristics and it is not specified whether Router 101 has Cone or AS NAT characteristics>
FIG. 13A is a diagram showing details of the connection sequence 5 shown in FIG. 7, and FIG. 13B is a diagram showing the port numbers shown in FIG. 13A.

  In the information exchange phase, the terminal 100 acquires NAT characteristic information including that the router 201 is PS NAT, the value of the port allocation interval α of the router 201, and the port number [GP24] of the router 201.

  When the router 101 has either Cone or AS NAT characteristics, the communication control unit 2007 of the terminal 200 causes the destination port number to include each port number included in the range of [GP12] or more and less than [GP14] for the above-described reason. It is considered that n P2P start request packets to be set are transmitted. Since the router 201 has PS NAT characteristics, each time the terminal 200 transmits a P2P start request packet, the router 201 assigns a port number incremented by the port assignment interval α to each of the packets. That is, the router 201 converts each transmission source port number of the P2P start request packet into [GP24 + α] to [GP24 + nα]. For this reason, the communication control unit 1007 of the terminal 100 transmits n P2P start request packets in which [GP24 + α] to [GP24 + nα] are set as destination port numbers (step S1701). At this time, the router 101 assigns the port number [GP13] prepared in the preparation phase to all P2P start request packets.

  At the time when the P2P start request packet is transmitted in step S1701, the router 201 does not set NAT so as to transfer to the terminal 200 packets whose destination port numbers are set to [GP24 + α] to [GP24 + nα]. Therefore, the router 201 discards n P2P start request packets in which [GP24 + α] to [GP24 + nα] are set as destination port numbers, and does not transfer the packets to the terminal 200. On the other hand, the router 101 transfers (n) P2P start request packets to set (IPG1, GP13) as the destination and set any one of [GP24 + α] to [GP24 + nα] as the source port number. The NAT is set so as to transfer the received packet to the terminal 100 (IPL1, LP1).

  On the other hand, the terminal 200 determines that the router 101 has either Cone or AS NAT characteristics in the information exchange phase, and the port numbers [GP12] and [GP14] opened by the terminal 100 in the NAT characteristics investigation phase and the preparation phase. NAT characteristic information including

  The communication control unit 2007 of the terminal 200 includes n P2P start request packets in which [LP1] is set as the destination port number and the value of [GP12] or more and less than [GP14] is set as the source port number (however, , N = [GP14] − [GP12]) (step S1702).

  Of the n P2P start request packets transmitted from the terminal 200, the router 101 transfers only the packet in which the port number [GP13] is set as the destination port number by the NAT of the router 201 to the terminal 100. When the terminal 100 receives the P2P start request packet transferred by the router 101, the terminal 100 transmits a P2P start response packet to the terminal 200 (step S1703).

  If the router 201 has PS NAT characteristics and the router 101 does not specify whether it has Cone or AS NAT characteristics, a P2P communication path is established between the terminals 100 and 200 according to the above-described processing. be able to.

  In this embodiment, the terminal 100 transmits n (that is, [GP14]-[GP12]) P2P start request packets, but the number of P2P start request packets transmitted by the terminal 100 is The number may not be n. For example, in the NAT characteristic investigation phase, after the router 201 opens the port [GP24], while the terminals 100 and 200 are exchanging NAT information with each other, other applications on the terminal 200 are not less than [GP24 + α]. There is a possibility of occupying some of a plurality of ports in the range of [GP24 + nα] or less. Considering this possibility, in order to increase the certainty of establishing the P2P communication path, the terminal 100 sets (n + m) pieces of port numbers [GP24 + α] to [GP24 + (n + m) α] of the router 201 as destination port numbers. P2P start request packets (where m is an integer) may be transmitted.

<Connection Sequence 6: Connection Sequence when Both Routers 101 and 201 have PS NAT Characteristics>
FIG. 14A is a diagram showing details of the connection sequence 6 shown in FIG. 7, and FIG. 14B is a diagram showing the port numbers shown in FIG. 14A.

  In the information exchange phase, the terminal 100 acquires NAT characteristic information including that the router 201 is PS NAT, the value of the port allocation interval α2 of the router 201, and the port number [GP24] of the router 201.

  PS NAT increments the port allocation interval α every time a port number is allocated to a packet. Therefore, the terminal 100 can predict that the port number used by the terminal 200 for P2P communication is [GP24 + α2].

  Therefore, the communication control unit 1007 transmits a P2P start request packet in which [GP24 + α2] is set as the destination port number and [LP1] is set as the source port number (step S1801). Since the router 101 also has PS NAT characteristics, the router 101 assigns a port number [GP14 + α1] to the P2P start request packet (where α1 is the port assignment interval of the router 101).

  At the point in time when the P2P start request packet is transmitted in step S1801, the router 201 does not set NAT so as to transfer the packet in which [GP24 + α2] is set to the destination port number to the terminal 200. Therefore, the router 201 discards the P2P start request packet and does not transfer the packet to the terminal 200. On the other hand, the router 101 sets NAT so as to transfer a packet in which (IPG1, GP14 + α1) is set as the destination and [GP24 + α2] is set as the source port number to the terminal 100 (IPL1, LP1).

On the other hand, the terminal 200, the information exchange phase, are acquired and that the router 101 has the PS NAT characteristic, the value of the port assignment interval [alpha] 1, the NAT characteristic information including the port number [GP 14]. The terminal 200 predicts that the port number used by the router 101 for P2P communication is [GP14 + α1], sets [GP14 + α1] as the destination port number, and sets [LP1] as the source port number. A packet is transmitted (step S1802).

  The router 101 transfers the P2P start request packet transmitted from the terminal 200 to the terminal 100. When the terminal 100 receives the P2P start request packet transferred by the router 101, the terminal 100 transmits a P2P start response packet to the terminal 200 (step S1803).

  When both the routers 101 and 201 have PS NAT characteristics, a P2P communication path can be established between the terminals 100 and 200 according to the above-described processing.

  In this embodiment, terminal 100 transmits one P2P start request packet with [GP24 + α2] set as the destination port number, but terminal 100 transmits a plurality of P2P start request packets. May be. In order to increase the possibility of successfully establishing the P2P communication path, the terminal 100 has a plurality of port numbers incremented by α2 as destination port numbers, such as [GP24 + 2 × α2] and [GP24 + 3 × α2], for example. The P2P start request packet may be transmitted.

  In this embodiment, a router is shown in which the assigned port number increases each time a port is assigned to a packet. However, the present invention is a router in which the assigned port decreases every time a port is assigned to a packet. It goes without saying that the same applies to the above. In this case, the port number allocation interval is obtained by calculating the absolute value of the difference between the two port numbers.

(Other variations)
The communication system according to the above-described embodiment may be configured as follows.

  The terminals 100 and 200 store the NAT characteristics determined when the P2P communication path is first established, and reuse the stored NAT characteristics to establish the P2P communication path for the second and subsequent times. The NAT characteristic investigation phase may be omitted. The terminals 100 and 200 may advance the P2P communication path establishment process without executing the NAT investigation phase process, and may execute the NAT characteristic investigation phase process only when the P2P communication path establishment fails. .

  According to such a configuration, the number of times the terminals 100 and 200 investigate the NAT characteristic is reduced, so that the second and subsequent P2P communication path establishment processes can be efficiently executed.

Further, when the terminals 100 and 200 store the NAT characteristics of the router once investigated, the terminals 100 and 200 may confirm that the stored NAT characteristics are not changed according to the following procedure. . Each of the terminals 100 and 200 transmits a packet including the port number of the connection destination apparatus to the connection destination apparatus during the period of the P2P communication path establishment phase or during the period when the terminals 100 and 200 are performing P2P communication. Also, server 001, a packet including the port number that the router 101 is used for communication with the server 001, and transmits to each terminal 100 and 200. The terminals 100 and 200 can confirm the NAT characteristic by comparing the port number notified from the connection destination device with the port number of the router used for communication with the server 001.

  Furthermore, when the terminal devices 100 and 200 determine that the router to which they are connected has either the Cone or AS NAT characteristic in the NAT characteristic investigation phase, whether the terminal apparatus 100 or 200 has the Cone NAT characteristic or the AS NAT characteristic. You may specify. Hereinafter, in the P2P communication path establishment phase, a process for each of the terminals 100 and 200 to determine whether its own router has the Cone or AS NAT characteristics will be described.

  FIG. 15A is a connection sequence diagram including processing for specifying whether the NAT characteristic of the router 101 is a Cone or AS NAT characteristic, and FIG. 15B is a diagram illustrating the port numbers shown in FIG. 15A.

  In the information exchange phase, the terminal 100 acquires the NAT characteristic information including the fact that the router 201 has the Cone NAT characteristic and the port number [GP22] of the router 201. Since the router 201 has the Cone NAT characteristic, the terminal 100 can predict that the port number of the router 201 that the terminal 200 uses for P2P communication is [GP22].

  Therefore, the communication control unit 1007 transmits, as a P2P start request, a packet in which [LP1] is set as the source port number and [GP22] is set as the destination port number (step S2001). The router 101 assigns a port number [GP13] to the packet transmitted from the terminal 100.

  At the stage where the P2P start request packet is transmitted in step S2001, the router 201 has set the NAT so that the router 201 forwards the packet in which the destination port number transmitted from the terminal 100 is set to [GP22]. Absent. Therefore, the router 201 discards the received P2P start request packet and does not transfer the packet to the terminal 200. At this stage, the router 101 sends a packet in which the port number [GP22] of the router 201 is set as the source port number and (IPG1, GP13) is set as the destination IP address and the destination port number to the terminal 100 ( Set NAT to forward to IPL1, LP1).

  On the other hand, in the information exchange phase, the terminal 200 has information indicating that the router 101 to which the terminal 100 is connected has either Cone or AS NAT characteristics, and the router 101 has opened in the NAT characteristics investigation phase and the preparation phase. NAT characteristic information indicating the port numbers [GP12] and [GP14] is acquired.

  When the terminal 200 obtains information on the NAT characteristics of the router 101, the terminal port number is set to [LP1], and the destination port number is set to a plurality of P2Ps set to a value between [GP12] and less than [GP14]. A start request packet is transmitted (step S2002).

  Of the plurality of P2P start request packets transmitted from the terminal 200, the router 101 transfers only one packet set to [GP13] as the destination port number by the router 201 to the terminal 100. When the terminal 100 receives the P2P start request packet transferred by the router 101, the terminal 100 transmits a P2P start response packet to the terminal 200 (step S2003).

  When the router 101 has either the Cone or AS NAT characteristics and the router 201 has the Cone NAT characteristics, a P2P communication path may be established between the terminals 100 and 200 according to the above-described processing. it can.

  Next, when receiving the P2P start response packet from the terminal 100, the communication control unit 2007 of the terminal 200 extracts the source port number of the packet, that is, the port number [GP13] of the router 101, and extracts the extracted port number [GP13 ] Is transmitted to the terminal 100 (step S2004).

  The relay characteristic determining unit 1004 of the terminal 100 refers to the notification packet transmitted from the terminal 200 and compares the port number [GP13] of the router 101 with the port number [GP12] investigated in the NAT characteristic investigation phase ( Step S2005). As a result of the comparison, if [GP13] is equal to [GP12] (Yes in step S2005), the relay characteristic determination unit 1004 determines that the router 101 has the Cone NAT characteristic (step S2006). In other cases (step S2005), the relay characteristic determining unit 1004 determines that the router 101 has the AS NAT characteristic (step S2006).

  Next, the relay characteristic determining unit 1004 stores the difference between the port numbers [GP13] and [GP12] in the predetermined storage unit as the port allocation interval Δ of the router 101 (step S2008).

  FIG. 16 is a connection sequence diagram showing processing when the NAT characteristics once investigated are stored.

  Since the terminal 100 has already investigated the router 101, the process in the NAT characteristic investigation phase is omitted, and the process in the preparation phase is executed.

<Preparation phase>
First, in the terminal 100, the address information investigation request transmission unit 1001 has the port number [LP1] of the terminal 100 set as the source port number and the port number [SP0] of the server 001 set as the destination port number. A port number investigation request packet is transmitted (step S2101).

  In the server 001, the address information investigation unit 0011 transmits the port number investigation response packet including the source port number of the port number investigation request packet, that is, the port number [GP11] of the router 101, via the communication unit 0014 ( Step S2102).

  Next, in the terminal 100, the direct communication preparation request transmission unit 1003 is used for P2P communication with the terminal 200 by transmitting a P2P communication preparation packet in which the IP address and port number of the terminal 200 are set as a destination. A port number [GP12] is prepared on the router 101 (step S2103). The router 101 sets NAT so that the packet transmitted from the terminal 200 to the router 101 (IPG1, GP12) can be received.

  Similarly, the terminal 200 omits the process in the NAT characteristic investigation phase and executes the process in the preparation phase (steps S2104 to S2106). The router 201 sets NAT so that the packet transmitted from the terminal 100 to the router 201 (IPG1, GP22) can be accepted.

<Information exchange phase>
Next, the terminal 200 transmits a NAT information notification packet including information on the NAT characteristics of the router 201 to the server 001 (step S2107). The NAT information includes different information as described below depending on whether or not the router 201 has the Port Reuse characteristic.
(E) When the router 201 does not have the Port Reuse characteristic, the NAT information includes at least the NAT characteristic of the router 201 and the port number [GP21] of the router 201. If it is determined that the router 201 has AS NAT characteristics during the previous P2P communication, the NAT information further includes the value of the port allocation interval Δ stored by the relay characteristic determination unit 2004.
(F) When the router 201 has the Port Reuse characteristic, the NAT information includes at least a port number [LP4] that the terminal 200 plans to newly open in the next P2P communication path establishment phase.

  Next, in the server 100, when the relay information transfer unit 0013 receives the NAT information notification packet transmitted from the terminal 200, the relay information transfer unit 0013 transfers the received NAT information notification packet to the terminal 100 via the communication unit 0014 (Step S100). S2108).

Similarly, in terminal 100, relay information transmission unit 1005 transmits a NAT information notification packet including information on the NAT characteristics of router 101 to server 001 (step S2109). The NAT information includes different information as described below depending on whether or not the router 101 has the Port Reuse characteristic.
(G) When the router 101 does not have the Port Reuse characteristic, the NAT information includes at least the NAT characteristic of the router 101 and the port number [GP11] of the router 101. If it is determined that the router 101 has AS NAT characteristics during the previous P2P communication, the NAT information further includes the value of the port allocation interval Δ stored by the relay characteristic determination unit 1004.
(H) When the router 101 has the Port Reuse characteristic, the NAT information includes at least the port number [LP3] that the terminal 100 plans to open newly in the next P2P communication path establishment phase.

  Next, in the server 100, when receiving the NAT information notification packet transmitted from the terminal 100, the relay information transfer unit 0013 transfers the received NAT information notification packet to the terminal 200 via the communication unit 0014 (step S100). S2110).

  In this embodiment, the terminal 200 transmits the NAT information notification packet before the terminal 100. However, the terminal 100 may transmit the NAT information notification packet before the terminal 200.

<Connection phase>
In the following, for simplification of description, it is assumed that the router 101 has the Cone NAT characteristic and the router 201 has the AS NAT characteristic.

  The communication control unit 1007 of the terminal 100 determines the terminal 200 based on the NAT information notification packet including the router 201 having AS NAT characteristics, the port number [GP21] of the router 201, and the port allocation interval Δ of the router 201. Predicts that the port number used by the router when the P2P communication with the terminal 100 is [GP21 + Δ]. The communication control unit 1007 transmits a P2P start request packet in which [GP21 + Δ] is set as the destination port number to the router 201 via the communication unit 1008 (step S2111). The router 101 assigns a port number [GP12] to the P2P start request packet.

  At the time when the P2P start request packet is transmitted in step S2111, the router 201 has not set NAT so as to transfer the packet in which [GP21 + Δ] is set to the destination port number to the terminal 200. Therefore, the router 201 discards the received P2P start request packet and does not transfer the packet to the terminal 200. On the other hand, the router 101 forwards the P2P start request packet to transfer the packet in which (IPG1, GP12) is set as the destination and [GP21 + Δ] is set as the source port number to the terminal 100 (IPL1, LP1). Set NAT to forward to.

  On the other hand, in the information exchange phase, the terminal 200 acquires NAT characteristic information including that the router 101 has the Cone NAT characteristic and the port number [GP11] of the router 101. The terminal 200 predicts that the port number used by the router 101 for P2P communication is [GP11], the P2P start request in which [GP11] is set as the destination port number and [LP1] is set as the source port number A packet is transmitted (step S2112). Since the router 101 has a Cone NAT characteristic, [GP12] is the same as [GP11].

  The router 101 transfers the P2P start request packet transmitted from the terminal 200 to the terminal 100. When the terminal 100 receives the P2P start request packet transferred by the router 101, the terminal 100 transmits a P2P start response packet to the terminal 200 (step S2113).

  In the P2P communication path establishment phase (steps S211 to S2113), when the terminals 100 and 200 fail to establish the P2P communication path, each of the terminals 100 and 200 may repeat the connection process again from the NAT characteristic investigation phase. .

  Further, in the flowchart shown in FIG. 7, one of the two terminals is connected to itself only when the router connected to the other terminal has PS NAT characteristics (steps S1204 to S1206). Although the NAT information notification packet information characteristic of the router is determined, the NAT characteristic of the router connected to itself may always be determined. In this case, each of the two terminals may select a connection sequence for each combination of NAT characteristics.

  In the present embodiment, for simplicity of explanation, the example in which the terminal 100 and the server 001 are connected by one router 101 is mainly shown. However, the terminal 100 and the server 001 include a plurality of routers. It may be connected via. The number of routers arranged between the terminal 200 and the server 001 is the same. This point will be described below.

  FIG. 17A is a diagram illustrating an example in which a terminal is connected to the global network via three routers.

  As shown in FIG. 17A, the terminal 100 is connected to the global network 00 via routers 1 to 3 having Cone NAT characteristics. When the terminal 100 transmits a packet in which [LP1] is set as the source port number, the packet reaches the network 00 through the port RP11 of the router 1, the port RP21 of the router 2, and the port RP31 of the router 3. . In this case, the server connected to the global network 00 can be regarded as if the terminal 100 is connected to itself through a router having one Cone NAT characteristic.

  FIG. 17B is a diagram illustrating another example in which the terminal is connected to the global network via three routers.

  As illustrated in FIG. 17B, the terminal 100 is connected to the global network 00 via the router 1 having the Cone NAT characteristic, the router 2 having the PS NAT characteristic, and the router 3 having the Cone NAT characteristic. When the terminal 100 transmits a packet in which [LP1] is set as the transmission source port number, the packet reaches the router 2 through the port PR11 of the router 1. Since the router 2 has PS NAT characteristics, different ports RP21, RP22, and RP23 are assigned to the received packet, and the packet is transferred to the router 3. Even when the router 3 has the Cone NAT characteristic, each packet having a different source port number is forwarded to the global network through a different port of the router 3, so that the server connected to the global network 00 is apparent. In addition, it can be considered that the terminal 100 is connected to itself via a single router having the PS NAT characteristic.

  In this way, even when the terminal and the server are connected via a plurality of routers, P2P communication is performed between the two terminals by selecting the connection sequence based on the apparent NAT characteristics that can be seen from the global network 00. A road can be established.

  Note that the functional blocks (FIGS. 2A and 2B) of the communication apparatus according to the above embodiment are programs that can cause a computer to execute the above-described processing procedure stored in a storage device (ROM, RAM, hard disk, etc.). It can also be realized by causing a computer to execute it. This computer is a concept including a computer incorporated in a portable terminal device. In this case, the program may be executed after being stored in the storage device via the recording medium, or may be executed directly from the storage medium. Further, the functional blocks of the communication device according to the above-described embodiment may be realized as an integrated circuit.

  The communication device, the communication method, and the communication system according to the present invention can investigate the NAT characteristic of the router by communicating with only one server having one IP address. Therefore, the present invention is useful for home appliances, communication devices, and the like that perform P2P communication between two terminals belonging to different private networks, for example.

1 is a block diagram showing the overall configuration of a communication system according to an embodiment of the present invention. The block diagram which shows schematic structure of the terminal shown by FIG. The block diagram which shows schematic structure of the terminal shown by FIG. The block diagram which shows schematic structure of the server shown by FIG. The sequence diagram which shows the outline of the process which the communication system which concerns on this embodiment performs in order to establish a P2P communication path Sequence diagram showing details of processing in the NAT characteristic investigation phase and the preparation phase shown in FIG. The flowchart which shows the detail of a process of step S908 shown by FIG. 4A. Figure 4 diagram for illustrating the step S910 shown in B Figure 4 diagram for illustrating the step S912 shown in B Figure 4 diagram for illustrating the step S913 shown in B Figure 4 diagram for illustrating the step S913 shown in B Sequence diagram showing details of processing in information exchange phase shown in FIG. The flowchart which shows the process which a terminal performs in the P2P communication path establishment phase shown by FIG. The figure which shows the detail of the connection sequence 1 shown by FIG. The figure which shows the detail of the connection sequence 2 shown by FIG. The figure which shows the port number which is shown in Figure 9A Figure for explaining router port number allocation interval The figure which shows the detail of the connection sequence 3 shown by FIG. The figure which shows the detail of the connection sequence 4 shown by FIG. The figure which shows the port number shown by FIG. 12A The figure which shows the relationship of the port number between the terminal and the router The figure which shows the detail of the connection sequence 5 shown by FIG. The figure which shows the port number shown by FIG. 13A The figure which shows the detail of the connection sequence 6 shown by FIG. The figure which shows the port number shown by FIG. 14A Connection sequence diagram including processing for specifying whether the NAT characteristic of the router is Cone or AS NAT characteristic The figure which shows the port number shown by FIG. 15A Connection sequence diagram showing processing in the case where the once investigated NAT characteristic is stored The figure which shows the example where the terminal is connected to the global network through three routers The figure which shows the other example in which the terminal is connected to the global network via three routers A diagram for explaining NAT called Full Cone NAT A diagram for explaining NAT called Restricted Cone NAT (R NAT) A diagram for explaining a NAT characteristic called Port Restricted Cone NAT (PR NAT) A diagram for explaining NAT called Address Sensitive Symmetric NAT (AS NAT) A diagram for explaining NAT called Port Sensitive Symmetric NAT (PS NAT) The figure for demonstrating the outline of a Port Reuse characteristic The figure for demonstrating the function of the router which has a Port Reuse characteristic The figure which shows the outline of STUN-TestI The figure which shows the outline of STUN-TestII The figure which shows the outline of STUN-TestIII Flowchart for investigating NAT characteristics by STUN Sequence diagram showing a P2P communication path establishment method using STUN Sequence diagram showing a processing procedure when a P2P communication path establishment method using STUN is applied to two communication terminals connected via a router having a Symmetric NAT characteristic Sequence diagram showing a processing procedure when a P2P communication path establishment method using STUN is applied to two communication terminals connected via a router having a Symmetric NAT characteristic Sequence diagram showing a conventional P2P communication path establishment method

Explanation of symbols

00 Global network 01, 02 Private network 001 Server 100, 200 (Home) terminal 101, 201 Router 0011 Address information investigation unit 0012 Address information transmission unit 0013 Relay information transfer unit 1001, 2001 Address information investigation request transmission unit 1002, 2002 Address information Reception unit 1003, 2003 Direct communication preparation request transmission unit 1004, 2004 Relay characteristic determination unit 1005, 2005 Relay information transmission unit 1006, 2006 Relay information reception unit 1007, 2007 Communication control unit 1008, 2008 Communication unit

Claims (20)

A communication device connected to a server via a first relay device having a NAT (network address translation) function,
A first transmission process for transmitting at least two address information investigation request packets having the same source port number and different destination port numbers to the server, and the source port number in the first transmission process An address information investigation request transmission unit that executes, in an arbitrary order, second transmission processing for transmitting at least one address information investigation request packet different from the used source port number to the server;
Communication in which the source port number is the same as any one of the source port numbers used in the first and second transmission processes at a destination different from the server during the first and second transmission processes. A communication preparation request transmission unit for transmitting a preparation packet;
Address information reception for receiving, from the server, at least three address information packets including a relay port number converted from the source port address of the address information investigation request packet in the data portion by the NAT function of the first relay device. And
A communication apparatus comprising: a relay characteristic determining unit that determines a NAT characteristic of the first relay apparatus based on a relay port number included in the received address information packet. The address information investigation request transmission unit
In the first transmission process,
A first address information investigation request packet in which the first port number of the server is set as the destination port number and the first local port number is set as the source port number;
A second address information investigation request packet in which a second port number different from the first port number is set as a destination port number, and the first local port number is set as a source port number; Send
In the second transmission process,
A third address information investigation request packet in which the first port number is set as the destination port number and the second local port number different from the first local port number is set as the source port number; Send
The address information receiving unit
A first address information packet returned in response to the first address information investigation request packet and including a first relay port number;
A second address information packet returned in response to the second address information investigation request packet and including a second relay port number;
Receiving a third address information packet returned in response to the third address information investigation request packet and including a third relay port number;
The relay characteristic determining unit determines a NAT characteristic of the first relay device based on at least two of the first to third relay port numbers.
The communication apparatus according to claim 1. A connection destination communication device is connected to the server via a second relay device having a NAT function,
The communication device
A relay information transmitting unit that transmits a first relay information packet including information indicating the determined NAT characteristic to the server;
A relay information packet receiving unit that receives, from the server, a second relay information packet including information indicating the NAT characteristics of the second relay device;
Communication control for transmitting a start request packet to the connection destination communication device based on NAT characteristics of the first and second relay devices to request the connection destination communication device to start peer-to-peer communication. And further comprising
The communication apparatus according to claim 2. The relay information transmission unit, when the relay characteristic determination unit determines that the first and second relay port numbers do not match, the third relay port number and the first and second relays Transmitting the first relay information packet including the difference value of the port number in the data part,
The communication apparatus according to claim 3. The relay information transmitting unit determines that the relay characteristic determining unit determines that the first and second relay port numbers have the same value, and a difference value between the third relay port number and the same value. The first relay information packet including the second relay port number in the data portion is determined when the first relay information packet is determined to be 1.
The communication apparatus according to claim 3. The relay information transmitting unit determines that the relay characteristic determining unit determines that the first and second relay port numbers have the same value, and a difference value between the third relay port number and the same value. The first relay information packet including the third relay port number in the data portion when it is determined that is greater than 1.
The communication apparatus according to claim 3. The relay information transmitting unit includes a fourth local port number included in a data part when the relay characteristic determining unit determines that the first relay port number and the first local port number match. 1 relay information packet is transmitted,
The communication apparatus according to claim 3. The relay information transmitting unit transmits the first relay information packet including information specifying the determined NAT characteristic in a data part,
The communication apparatus according to claim 3. The second relay information packet includes at least one of information specifying the NAT characteristic of the second relay device, and the fourth to sixth relay port numbers and the fifth local port number notified from the server. In the data part,
The communication control unit selects one from a plurality of predetermined connection sequences based on information included in the second relay information packet.
The communication apparatus according to claim 3. In the second relay information packet, the fourth relay port number and the fifth relay port number match, and the difference value between the sixth relay port number and the matched value is 1 information indicating that, and said fifth relay port number includes the data unit,
The communication control unit transmits a start request packet in which the first local port number is set as a source port number and the fifth relay port number is set as a destination port number. ,
The communication apparatus according to claim 9. In the second relay information packet, the fourth relay port number and the fifth relay port number match, and the difference value between the sixth relay port number and the matched value is greater than 1 information indicating that, and said fifth relay port number and the sixth relay port number of including in the data unit,
In the communication control unit, a value in a range in which the first local port number is set as a source port number and is greater than or equal to the fifth relay port number and less than the sixth relay port number is a destination port number Transmitting a plurality of start request packets set to the connection destination communication device,
The communication apparatus according to claim 9. The second relay information packet includes information indicating that the second relay device uses the same port number as the local port number of the connection destination communication device, and the fifth local port number. Included in the department,
The communication control unit transmits a start request packet in which the first local port number is set as a source port number and the fifth local port number is set as a destination port number. ,
The communication apparatus according to claim 9. The second relay information packet includes information indicating that the fourth and fifth relay port numbers do not match, the difference between the sixth relay port number, and the fourth and fifth relay port numbers. Value in the data part,
The communication control unit determines that the relay characteristic determining unit determines that the first and second relay port numbers are the same value, and a difference value between the third relay port number and the same value is determined. 1 is determined, the first local port number is set as the source port number, and the sixth relay port number and the fourth and fifth relay ports are set as the destination port number. A start request packet in which a value obtained by adding a difference value of numbers is set is transmitted,
The communication apparatus according to claim 9. The second relay information packet includes information indicating that the fourth and fifth relay port numbers do not match, the difference between the sixth relay port number, and the fourth and fifth relay port numbers. Value in the data part,
The communication control unit determines that the relay characteristic determination unit has the same value as the first and second relay port numbers, and the difference value between the third relay port number and the same value is 1 When it is determined that the port number is greater than the first port number, the first local port number is set as the source port number, and the sixth relay port number is the destination port number. A plurality of start request packets each set with a value incremented by a difference value are transmitted,
The communication apparatus according to claim 9. The second relay information packet includes information indicating that the fourth and fifth relay port numbers do not match, the difference between the sixth relay port number, and the fourth and fifth relay port numbers. Value in the data part,
When the communication control unit determines that the first and second relay port numbers do not match, the first local port number is set as the source port number, and the destination port number A start request packet in which a value obtained by adding a sixth relay port number and a difference value between the fourth and fifth relay port numbers is set is transmitted.
The communication apparatus according to claim 8. The communication control unit transmits a start request packet in which the fourth local port number is set as a source port number based on a NAT characteristic of the first relay device.
The communication device according to claim 7. The communication preparation request transmitter transmits a TTL (Time) included in a header of the communication preparation packet so that the communication preparation packet reaches the first relay device and does not reach the second relay device. (To Live) is adjusted.
The communication apparatus according to claim 3. The relay characteristic determination unit holds the NAT characteristics of each of the first and second relay devices,
The communication control unit transmits the connection start request packet to the connection destination communication device based on the retained NAT characteristic when the next communication with the connection destination communication device is performed.
The communication apparatus according to claim 3. A communication method for communication by a communication terminal connected to a server via a first relay device having a NAT (network address translation) function,
A first transmission process for transmitting at least two address information investigation request packets having the same source port number and different destination port numbers to the server, and the source port number in the first transmission process A second transmission process of transmitting at least one address information investigation request packet different from the used source port number to the server in any order;
Communication in which the source port number is the same as any one of the source port numbers used in the first and second transmission processes at a destination different from the server during the first and second transmission processes. Send a prepare packet,
Receiving from the server at least three address information packets including a relay port number converted from a source port address of the address information investigation request packet in the data portion by the NAT function of the first relay device;
Determining a NAT characteristic of the first relay device based on a relay port number included in the received address information packet;
Communication method. A communication system,
Server,
A first relay device having a NAT (network address translation) function;
A second relay device having a NAT function;
A first communication device connected to the server via the first relay device;
A second communication device connected to the server via the second relay device,
Each of the first and second communication devices includes:
A first transmission process for transmitting at least two address information investigation request packets having the same source port number and different destination port numbers to the server, and the source port number in the first transmission process An address information investigation request transmission unit that executes, in an arbitrary order, a second transmission process of transmitting at least one address investigation information request packet different from the used source port number to the server;
Communication in which the source port number is the same as any one of the source port numbers used in the first and second transmission processes at a destination different from the server during the first and second transmission processes. A communication preparation request transmission unit for transmitting a preparation packet;
An address for receiving from the server at least three address information packets including the relay port number converted from the source port address of the address information investigation request packet in the data part by the NAT function of the relay device to which each is connected An information receiver;
A relay characteristic determining unit that determines a NAT characteristic of a relay device to which each is connected based on a relay port number included in the received address information packet;
The server
An address information investigation unit that extracts the relay port number included in the address information investigation request packet transmitted from each of the first and second communication devices;
An address information transmission unit that transmits an address information investigation request response packet including the extracted relay port number in a data part,
Communications system.

Images