CN110113439B - NAT traversal method - Google Patents

Abstract

The invention discloses a NAT (network Address translation) traversal method, which comprises two stages: the first stage is an NAT type detection stage, a client side carries out NAT type detection by means of a STUN server and sends information such as an NAT type, a host address, a mapping address and the like to an NAT traversal server; the second stage is NAT traversing stage, the NAT traversing server will determine the corresponding traversing strategy according to the NAT type detection result of the client of both communication sides, and send the strategy and the related information to the client, after the client receives the NAT traversing strategy, the client will proceed NAT traversing according to the strategy; and when the NAT cannot be traversed, forwarding by adopting a transit server. The method greatly improves the traversing success rate of the symmetric NAT.

Description

NAT traversal method

Technical Field

The invention relates to the technical field of computer network communication, in particular to a method for NAT traversal.

Background

The rapid development of the Internet and the limitation of the number of IPv4 addresses make network address translation (nat) devices widely used. In existing Internet networking environments, many devices are located behind NAT devices or firewalls. The NAT device allows a host within a private network to actively initiate a connection to a public network server, but prohibits an active connection in the opposite direction. In many cases, however, 2 devices communicating with each other may all be behind NAT devices and cannot communicate directly with a peer-to-peer (P2P) connection.

To address the NAT traversal problem, the industry has proposed many solutions. These technologies can be divided into two broad categories, one based on NAT devices and the other based on terminals. The terminal-based solutions include stun (session traffic utilization for nat) protocol, turn (traffic utilization Relay nat) protocol, and ice (interactive Connectivity establishment) framework. However, the solution based on the NAT device often needs to upgrade or modify the existing NAT device, and the NAT cascading problem cannot be effectively solved. In the existing solution based on the terminal, for the symmetric NAT, the STUN protocol cannot traverse, and the TURN protocol and the ICE framework are inefficient.

Disclosure of Invention

The invention aims to overcome the defects of the existing NAT traversal scheme, provides a new NAT traversal method and greatly improves the traversal success rate of the symmetric NAT.

The purpose of the invention can be achieved by adopting the following technical scheme:

a NAT traversing method is applied to a system formed by a server and a client, wherein the server is deployed on a public network and comprises an NAT traversing server, a STUN server and a transit server, the client comprises an NAT type detection module and an NAT traversing module, and the method comprises the following steps:

s1, NAT type detection stage: the client side carries out NAT type detection by means of the STUN server and sends a detection result to the NAT traversal server;

s2, NAT traversal stage: after the NAT traversal server receives the NAT type detection results of the two communication client sides, the NAT traversal server determines different NAT traversal strategies, and then the NAT traversal strategies and the NAT type, the host address and the mapping address of the opposite client side, the port increment delta P of the increment symmetric NAT, the maximum value P of the measurement port of the local random symmetric type or the global random symmetric type NAT are used_maxAnd a minimum value P_minSending the data to the client of the other party; after receiving the NAT traversal strategy and the related information, the client sides of the two parties carry out NAT traversal according to the received NAT traversal strategy; and if the NAT traversal fails, forwarding by adopting a transit server.

Furthermore, the NAT type comprises a blocking type, an open type, a firewall type, a full cone type, an IP limiting cone type, a port limiting cone type, an increment symmetry type, a local random symmetry type and a global random symmetry type; the STUN server comprises 1 IP address and 2 ports, wherein the 2 ports are respectively a port SP1 and a port SP 2.

Further, the step S1 is as follows:

s101, the IP address of the host of the client X is IP_{X_host}Binding Port of client X_X1Port SP1 of access STUN server, obtaining Port from response message returned from STUN server_X1Mapped address IP of access port SP1_{X1_reflex}:Port_{X1_SP1}；

S102, if the overtime does not return, the system is in a blocking type, and the step S112 is entered;

s103, if IP_{X1_reflex}Equal to IP_{X_host}Checking whether the door is open or firewall type, namely executing step S104, otherwise executing step S106;

s104, Port of binding client X_X1The number of ports is such that,the port SP1 of the STUN server is accessed again, the CHANGE port bit of the CHANGE-REQUEST attribute in the STUN command REQUEST message is set to be 1, and the STUN server is required to return a CHANGE port response through the other port SP 2;

s105, if a change port response returned by the STUN server through a different port SP2 is received, the result shows that no NAT is converted, namely the STUN is an open type, otherwise, the result shows that the STUN is a firewall type, and the step S112 is executed;

s106, Port of binding client X_X1A port SP1 of the STUN server is accessed again, the CHANGE port bit of the CHANGE-REQUEST attribute in the STUN command REQUEST message is set to be 1, and the STUN server is required to return a CHANGE port response through another port SP 2;

s107, Port of binding client X_X1Port, Port SP2 of access STUN server, obtain Port_X1Mapped Port Port of Access Port SP2_{X1_SP2}；

S108, if Port_{X1_SP1}Equal to Port_{X1_SP2}Explaining that the NAT is a cone NAT, step S109 is executed to further distinguish which cone NAT is the NAT; otherwise, if the NAT is a symmetric NAT, step S110 is executed to further distinguish which symmetric NAT is;

s109, if a change port response returned by the STUN server through the port SP2 is received, the NAT is described as a full cone type or an IP restriction cone type, otherwise, the NAT is described as a port restriction cone type, and the step S112 is executed;

s110, binding different ports of client X_X2Port, Port SP1 of access STUN server, obtain Port_X2Mapped Port Port of Access Port SP1_{X2_SP1}(ii) a Port of binding client_X2Port, Port SP2 of access STUN server, obtain Port_X2Mapped address IP of access port SP2_{X2_reflex}:Port_{X2_SP2}；

S111, compare Port_{X1_SP1},Port_{X1_SP2},Port_{X2_SP1}And Port_{X2_SP2}A relationship of four;

if Port_{X1_SP2}-Port_{X1_SP1}＝Port_{X2_SP1}-Port_{X1_SP2}＝Port_{X2_SP2}-Port_{X2_SP1}If the NAT is in an increment symmetric type, the port increment is delta P; if not, the maximum value P of the four ports is calculated as a random symmetry type_maxAnd a minimum value P_minIf P is_maxAnd P_minIf the difference value of (A) is less than R, judging the local random symmetry type; if the R is larger than the R, the system is in a global random symmetry type;

s112, the client side sends the NAT type and the host address IP_{X_host}:Port_X2And mapping address IP_{X2_reflex}:Port_{X2_SP2}Sending to NAT traversal server, wherein Port for open type, firewall type or cone type NAT_X2Get Port_X1Value of (2), IP_{X2_reflex}:Port_{X2_SP2}Mapping address IP_{X1_reflex}:Port_{X1_SP1}And (3) the detection process ends.

Further, in step S2, when the NAT type of the client a is a port restricted cone type or a firewall type, and the NAT type of the client B is an incremental symmetry type, the traversal process is as follows:

s201, client B binding Port_B3Port, sending request to Port SP1 of STUN server to obtain mapping Port_{B3_SP1}；

S202, binding Port by client B_B3Port, mapping address IP to client A_{A2_reflex}:Port_{A2_SP2}Sending a request in which IP_{A2_reflex}:Port_{A2_SP2}Mapping address IP corresponding to client A_{X2_reflex}:Port_{X2_SP2}The NAT passes through the server and is sent to the client B; client B then listens to Port_B3A port for waiting for a request message of a client A;

s203, binding Port by client B_B3Port, sending request to Port SP2 of STUN server to obtain mapping Port_{B3_SP2}(ii) a Client B will Port_{B3_SP1}、Port_{B3_SP2}And the port increment delta P obtained in the NAT type detection process_BSending the data to an NAT traversal server; the NAT traversal server forwards the received message to the client A, and the client A starts NAT traversal after receiving the message;

s204, binding Port by client A_A2Port mapping each predicted Port in Port prediction interval to client B in sequence_BpredSending a request message, wherein Port_A2Port corresponding to client A_X2Port, Port_Bpred＝Port_{B3_SP1}+k*ΔP_BK is a positive integer, and Port_{B3_SP1}<Port_Bpred<Port_{B3_SP2}；

S205, when the client B is at the Port_B3When receiving the request of the client A, showing that the NAT traversal is successful; mapping Port of client B to client A_{A2_SP2}And sending a response to end the NAT traversal process.

Further, in step S2, when the NAT type of the client a is a port-restricted cone type or a firewall type, and the NAT type of the client B is a local random symmetric type or a global random symmetric type, the traversing process is as follows:

s211, according to the success crossing probability P set by the system_sThe optimal parameters N and M are calculated by the following formula, where T is the total number of available ports, T65535 and 1024, and N and M are [1, T-]The calculated optimal parameters N and M can reduce the NAT traversal time;

s212, client B binds N different Port_B31、Port_B32……Port_B3NMapping Port to client A respectively_{A2_SP2}Sending requests, obtaining N different mapping ports in total, wherein the Port_{A2_SP2}Mapping Port corresponding to client A_{X2_SP2}The NAT passes through the server and is sent to the client B; then the client B monitors the N ports and waits for the request sent by the client A;

s213, client A binds local Port_A2Port, mapping Port Range to client B_{B2_SP2}-M/2+1,Port_{B2_SP2}+M/2]M ports within a Port send requests, where the Port sends the request_A2Port corresponding to client A_X2Port, Port_{B2_SP2}Mapping Port corresponding to client B_{X2_SP2}；

S214, when the client B receives the request of the client A at 1 monitoring port, the NAT traversal is successful; the client B maps the Port to the client A through the Port_{A2_SP2}And sending a response to end the NAT traversal process.

Further, in step S2, when the NAT types of the two clients are both incremental symmetric types, the traversal process is as follows:

s221, client B binding Port_B3Sending a request to Port SP1 of the STUN server to obtain a mapping Port_{B3_SP1}；

S222, client B binding Port_B3Mapping N predicted ports within a Port prediction interval to client A sequentially_ApredSending a request; then monitoring the ports, and waiting for a request message sent by the client A; wherein, Port_Apred＝Port_{A2_SP2}+(k+Offset)*ΔP_AK is [1, N]Positive integers in the range, Offset is a natural number more than or equal to 0, and is used for avoiding an invalid port of the estimated NAT equipment of the client A; port (Port)_{A2_SP2}Mapping Port corresponding to client A_{X2_SP2}，ΔP_AThe parameters are port increment of NAT equipment of the client A and are sent to the client B through an NAT traversal server;

s223, client B binding Port_B3Sending a request to Port SP2 of the STUN server to obtain a mapping Port_{B3_SP2}Client B will Port_{B3_SP1}N, Offset and Port increment Δ P_BSending the NAT traversal request to an NAT traversal server, forwarding the NAT traversal request to a client A by the NAT traversal server, and starting NAT traversal after the client A receives the NAT traversal request;

s224, client A binding Port_A3Sending a request to Port SP2 of the STUN server to obtain a mapping Port_{A3_SP2}(ii) a Determining Port_{A3_SP2}Whether or not to be greater than or equal to Port_{A2_SP2}+Offset*ΔP_AIf the condition is not satisfied, the client A updates the local Port_A3For other different values, the request is sent to Port SP2 of the STUN server again to update the mapping Port_{A3_SP2}Repeating the above process until the mapping port satisfies the above condition; to reduce the impact of other processes or other clients on NAT device port assignment, client B's 1 st predicted port is calculated according to the following formula:

s225, client A binding Port_A3Mapping N predicted ports within a Port prediction interval to client B sequentially_BpredSending a request in which Port_Bpred＝Port_{Bpred_1}+2*k*ΔP_BK is [0, N-1 ]]A natural number within the range;

s226, when the client B receives the request of the client A at 1 monitoring port, indicating that the NAT traversal is successful; and the client B sends a response to the client A to finish the NAT traversal process.

Further, in step S2, when the NAT type of the client a is a local random symmetric type or a global random symmetric type, and the NAT type of the client B is an incremental symmetric type, the traversal process is as follows:

s231, client B binding Port_B3Port, sending request to Port SP1 of STUN server to obtain mapping Port_{B3_SP1}；

S232, binding N different local ports by the client B, and sequentially mapping the ports to the predicted mapping ports of the client A_ApredSending a request in which Port_ApredIs (P)_max,P_min) An integer randomly selected within the range;

s233, binding Port to client B_B3Port, sending request to Port SP2 of STUN server to obtain mapping Port_{B3_SP2}(ii) a Client B will Port_{B3_SP1}And Port_{B3_SP2}Sending the NAT traversal request to an NAT traversal server, forwarding the NAT traversal request to a client A by the NAT traversal server, and starting NAT traversal after the client A receives the NAT traversal request;

s234, client A binds local Port_A3Mapping each predicted Port within a Port prediction interval to client B sequentially_BpredSending a request message, whichMiddle Port_Bpred＝Port_{B3_SP1}+k*ΔP_BK is a positive integer, and Port_{B3_SP1}<Port_Bpred<Port_{B3_SP2}；

S235, when the client B receives the request of the client A, showing that the NAT traversal is successful; the client B sends a response to the client A, and the NAT traversal process is ended;

s236, if the client A does not receive the response of the client B after time-out, the local Port is updated_A3And returning to the step S234 for retries for other different values, and retrying for Try times in total.

Further, in step S2, when the mapping IP addresses IP of both clients are used_{X2_reflex}And the host IP addresses and the corresponding mapping IP addresses of the two client sides are different, the two client sides are judged to be under the same NAT equipment, and the traversing process is as follows:

s241, the host address IP from the client A to the client B_{B_host}:Port_B2Sending a request, IP_{B_host}:Port_B2Host address IP corresponding to client B_{X_host}:Port_X2The NAT passes through the server and is sent to the client A;

s242, the client B receives the request packet sent by the client A and obtains a source address of the request packet; after obtaining the source address of the request packet, the client B sends a request to the address, and then the process goes to step S243; if the client B does not receive the request packet after timeout, go to step S244;

s243, after receiving the request of the client B, the client A sends a response to the client B; if the client B receives the response, the direct traversing process is finished;

s244, the host address IP from the client B to the client A_{A_host}:Port_A2Sending a request, IP_{A_host}:Port_A2Host address IP corresponding to client A_{X_host}:Port_X2The NAT passes through the server and is sent to the client B;

s245, the client A receives the request packet sent by the client B and obtains the source address of the request packet; after obtaining the source address of the request packet, the client a sends a request to the address, and then the process goes to step S246; if the client A does not receive the request packet after overtime, the direct NAT traversal fails, and the clients of the two parties can be regarded as different NAT devices and try again to perform NAT traversal by adopting a corresponding traversal strategy according to the NAT type;

s246, after receiving the request of the client A, the client B sends a response to the client A; and if the client A receives the response, ending the direct NAT traversal process.

Further, in step S2, when the NAT types of the clients are both port restricted cone type or firewall type, the traversing process is as follows:

s251, client A binding Port_A2Mapping an address IP to client B_{B2_reflex}:Port_{B2_SP2}Sending a request in which Port_A2Port corresponding to client A_X2Port, IP_{B2_reflex}:Port_{B2_SP2}Mapping address IP corresponding to client B_{X2_reflex}:Port_{X2_SP2}The NAT passes through the server and is sent to the client A; after the client A sends a request, monitoring the port and waiting for the request of the client B;

s252, client B binding Port_B2Mapping an address IP to a client A_{A2_reflex}:Port_{A2_SP2}Sending a request in which Port_B2Port corresponding to client B_X2Port, IP_{A2_reflex}:Port_{A2_SP2}Mapping address IP corresponding to client A_{X2_reflex}:Port_{X2_SP2}The NAT passes through the server and is sent to the client B; if the client A does not receive the request of the client B after overtime, the NAT traversal fails;

s253, after receiving the request of the client B, the client A sends a response to the client B, and after receiving the response, the client B finishes the NAT traversal process; if the client B does not receive the response after time out, the NAT traversal fails.

Further, in step S2, when the NAT type of the client a is other than the blocking type, the client B is the open type, and there is no NAT device, the traversal process is as follows:

s261, mapping address IP of client A to B_{B2_reflex}:Port_{B2_SP2}Sending a request in which IP_{B2_reflex}:Port_{B2_SP2}Mapping address IP corresponding to client B_{X2_reflex}:Port_{X2_SP2}(ii) a After the client A sends a request, monitoring the port and waiting for the request of the client B;

s262, the client B receives the request packet sent by the client A and obtains the source address of the request packet; after obtaining the source address of the request packet, the client B sends a request to the address; if the client B does not receive the request packet after overtime, the NAT traversal fails;

s263, after receiving the request of the client B, the client A sends a response to the client B; and after receiving the response, the client B ends the NAT traversal process.

Further, in step S2, when the NAT type of the client a is full cone type or IP restricted cone type, and the NAT type of the client B is other types except blocking type, the traversal process is as follows:

s271, client A binding Port_A2Mapping IP address IP to B_{B2_reflex}Any 1 Port on, where Port sends the request_A2Port corresponding to client A_X2Port, IP_{B2_reflex}Mapping IP address IP corresponding to client B_{X2_reflex}(ii) a After the client A sends a request, monitoring the port and waiting for the request of the client B;

s272, mapping address IP from client B to client A_{A2_reflex}:Port_{A2_SP2}Sending a request in which IP_{A2_reflex}:Port_{A2_SP2}Mapping address IP corresponding to client A_{X2_reflex}:Port_{X2_SP2}(ii) a If the client A does not receive the request of the client B after overtime, the NAT traversal fails;

s273, after receiving the request of the client B, the client A sends a response to the client B; and after the client B receives the response, the traversing process is ended.

Compared with the prior art, the invention has the following advantages and effects:

1. the method greatly improves the success rate of the symmetric NAT penetration, in particular to the success rate under the scene of port limited conical or firewall type-increment symmetric type, port limited conical or firewall type-random symmetric type, increment symmetric type-increment symmetric type and random symmetric type-increment symmetric type;

2. according to different network topologies and NAT equipment types, different NAT traversal modes are adopted, so that the method has stronger adaptability;

3. different from the NAT type detection flow in the traditional STUN protocol, the STUN server only needs 1 IP address in the method.

Drawings

Fig. 1 is a flowchart of a NAT traversal method according to an embodiment of the present invention;

fig. 2 is a schematic system structure diagram of a NAT traversal method according to an embodiment of the present invention;

FIG. 3 is a flow diagram of NAT type probing;

fig. 4 is a schematic diagram of port restricted cone or firewall-incremental symmetric NAT traversal;

FIG. 5 is a schematic diagram of port restricted cone or firewall-random symmetric NAT traversal;

fig. 6 is a schematic diagram of incremental symmetric-incremental symmetric NAT traversal;

fig. 7 is a schematic diagram of random-incremental symmetric NAT traversal.

Detailed Description

In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

Examples

As shown in fig. 1, the present embodiment discloses a NAT traversal method, and as shown in fig. 2, an application system of the method includes a server and a client; the server comprises three modules of an NAT traversing server, an STUN server and a transfer server; the client comprises an NAT type detection module and an NAT traversal module; the server is deployed in the public network.

As shown in fig. 1, a flowchart of a Network Address Translation (NAT) traversal method according to an embodiment of the present invention is shown. The method comprises the following steps:

step S1, NAT type detection stage: at this stage, the client performs NAT type detection by means of the STUN server and sends the detection result to the NAT traversal server.

Step S2, NAT traversal stage: after the NAT traversal server receives the NAT type detection results of the two communication client sides, the NAT traversal server can determine different NAT traversal strategies according to the information, and then the NAT traversal strategies and the NAT type, the host address and the mapping address of the opposite client side, the port increment delta P of the increment symmetric type NAT and the maximum value P of the measurement port of the local random symmetric type or global random symmetric type NAT are used_maxAnd minimum value P_minSending the data to the client of the other party; after receiving the NAT traversal strategy and the related information, the client sides of the two parties carry out NAT traversal according to the strategy; and if the NAT traversal fails, forwarding by adopting a transit server.

In step S1, the client needs to probe the NAT type; in the NAT type detection flow of the method, the STUN server only needs 1 IP address unlike the standard STUN protocol; the detected NAT types comprise a blocking type, an opening type, a firewall type, a full cone type or an IP limiting cone type, a port limiting cone type, an increment symmetry type, a local random symmetry type and a global random symmetry type, wherein the full cone type and the IP limiting cone type are not distinguished in the detection method;

different from the traditional detection process, in the NAT type detection process of the method, the REQUEST sent by the client to the STUN server does not require the STUN server to CHANGE the IP address for sending the response, and the CHANGE IP bit of the CHANGE-REQUEST attribute is set to be 0; the 2 ports of the STUN server are respectively a port SP1 and a port SP2, in the embodiment of the invention, SP1 is 3478, and SP2 is 3479; the specific steps are shown in fig. 3, and include:

step S101, the IP address of the host of the client X is IP_{X_host}Binding Port of client X_X1Port, access SThe 3478 Port of the TUN server obtains the Port from the MAPPE-ADDRESS attribute of the response message returned by the STUN server_X1Accessing mapped address IP of 3478 port_{X1_reflex}:Port_{X1_3478}；

Step S102, if overtime does not return, the system is in a blocking type, and the process goes to step S112;

step S103, if IP_{X1_reflex}Equal to IP_{X_host}Checking whether the door is open or firewall type, namely executing step S104, otherwise executing step S106;

step S104, binding Port of client X_X1A port, which accesses the 3478 port of the STUN server again, wherein the CHANGE port bit of the CHANGE-REQUEST attribute in the STUN command REQUEST message is set to 1, and the STUN server is required to return a CHANGE port response through another port 3479;

step S105, if a change port response returned by the STUN server through a different port 3479 is received, it indicates that no NAT is performed, i.e. the STUN server is an open type, otherwise, the STUN server is a firewall type, and step S112 is executed;

step S106, binding Port of client X_X1A port, which accesses the 3478 port of the STUN server again, wherein the CHANGE port bit of the CHANGE-REQUEST attribute in the STUN command REQUEST message is set to 1, and the STUN server is required to return a CHANGE port response through another port 3479;

step S107, Port of binding client X_X1Port, 3479 Port accessing STUN server, obtaining Port_X1Accessing a mapped Port of 3479 ports_{X1_3479}；

Step S108, if Port_{X1_3478}Equal to Port_{X1_3479}Explaining that the NAT is a cone type, step S109 is executed to further distinguish which cone type is used; otherwise, if the NAT is a symmetric NAT, step S110 is executed to further distinguish which symmetric NAT is;

step S109, if a change port response returned by the STUN server through the port 3479 is received, it indicates that the NAT is a full cone type or an IP restriction cone type, otherwise, it is a port restriction cone type, and step S112 is executed;

step S110, binding different Pors of client Xt_X2Port, 3478 Port of STUN Server Access, Port_X2Mapped Port Port of Access Port 3478_{X2_3478}(ii) a Port of binding client_X2Port, 3479 Port accessing STUN server, obtaining Port_X2Mapped address IP for Access Port 3479_{X2_reflex}:Port_{X2_3479}；

Step S111, comparing Port_{X1_3478},Port_{X1_3479},Port_{X2_3478}And Port_{X2_3479}A relationship of four;

if Port_{X1_3479}-Port_{X1_3478}＝Port_{X2_3478}-Port_{X1_3479}＝Port_{X2_3479}-Port_{X2_3478}If the NAT is in an increment symmetric type, the port increment is delta P; if not, the maximum value P of the four ports is calculated as a random symmetry type_maxAnd a minimum value P_minIf P is_maxAnd P_minIf the difference value of (A) is less than R, judging the local random symmetry type; if the R is larger than the R, the system is in a global random symmetry type; in the embodiment of the invention, R is 1000;

step S112, the client end sets the NAT type and the host address IP_{X_host}:Port_X2And mapping address IP_{X2_reflex}:Port_{X2_3479}Sending to NAT traversal server, wherein Port for open type, firewall type or cone type NAT_X2Get Port_X1Value of (2), IP_{X2_reflex}:Port_{X2_3479}Mapping address IP_{X1_reflex}:Port_{X1_3478}And (3) the detection process ends.

In step S2, that is, in the NAT traversal stage, when the NAT type of the client a is the port restriction cone type or the firewall type, and the NAT type of the client B is the incremental symmetry type, as shown in fig. 4, the traversal process is as follows:

step S201, client B binding Port_B3Port, sending request to 3478 Port of STUN server to obtain mapping Port_{B3_3478}；

Step S202, client B binding Port_B3Port, mapping address IP to client A_{A2_reflex}:Port_{A2_3479}A request is sent to the mobile station to request,wherein IP_{A2_reflex}:Port_{A2_3479}Mapping address IP corresponding to client A_{X2_reflex}:Port_{X2_3479}The NAT passes through the server and is sent to the client B; client B then listens to Port_B3A port for waiting for a request message of a client A;

step S203, binding Port by client B_B3Port for sending request to 3479 Port of STUN server to obtain mapped Port_{B3_3479}(ii) a Client B will Port_{B3_3478}、Port_{B3_3479}And the port increment delta P obtained in the NAT type detection process_BSending the data to an NAT traversal server; the NAT traversal server forwards the received message to the client A, and the client A starts NAT traversal after receiving the message;

step S204, binding Port by client A_A2Port mapping each predicted Port in Port prediction interval to client B in sequence_BpredSending a request message, wherein Port_A2Port corresponding to client A_X2Port, Port_Bpred＝Port_{B3_3478}+k*ΔP_BK is a positive integer, and Port_{B3_3478}<Port_Bpred<Port_{B3_3479}；

Step S205, when the client B is at the Port_B3When receiving the request of the client A, showing that the NAT traversal is successful; mapping Port of client B to client A_{A2_3479}And sending a response to finish the whole flow.

In step S2, when the NAT type of the client a is a port-restricted cone type or a firewall type, and the NAT type of the client B is a local random symmetric type or a global random symmetric type, as shown in fig. 5, the specific traversal process is as follows:

step S211, according to the success crossing probability P set by the system_sThe optimal parameters N and M are calculated by the following formula, where T is the total number of available ports, T65535 and 1024, and N and M are [1, T-]The calculated optimal parameters N and M can reduce the NAT traversal time; in the present embodiment, P_sN, calculated as 0.95, M are 427 and 448 respectively;

step S212, client B binds N different Port_B31、Port_B32……Port_B3NMapping Port to client A respectively_{A2_3479}Sending requests, obtaining N different mapping ports in total, wherein the Port_{A2_3479}Mapping Port corresponding to client A_{X2_3479}The NAT passes through the server and is sent to the client B; then the client B monitors the N ports and waits for the request sent by the client A;

step S213, client A binds local Port_A2Port, mapping Port Range to client B_{B2_3479}-M/2+1,Port_{B2_3479}+M/2]M ports within a Port send requests, where the Port sends the request_A2Port corresponding to client A_X2Port, Port_{B2_3479}Mapping Port corresponding to client B_{X2_3479}；

Step S214, when the client B receives the request of the client A at 1 monitoring port, it shows that the NAT traversal is successful; the client B maps the Port to the client A through the Port_{A2_3479}And sending a response, and finishing the traversing process.

In step S2, when the NAT types of the two clients are both incremental symmetric types, the process shown in fig. 6 is used to perform NAT traversal, and the specific steps are as follows:

step S221, client B binds Port_B3Sending a request to the 3478 Port of the STUN server to obtain a mapping Port_{B3_3478}。

Step S222, client B binding Port_B3Mapping N predicted ports within a Port prediction interval to client A sequentially_ApredSending a request; then monitoring the ports, and waiting for a request message sent by the client A; wherein, Port_Apred＝Port_{A2_3479}+(k+Offset)*ΔP_AK is [1, N]Positive integers in the range, Offset is a natural number more than or equal to 0, and is used for avoiding an invalid port of the estimated NAT equipment of the client A; port (Port)_{A2_3479}Mapping port corresponding to client APort_{X2_3479}，ΔP_AThe parameters are port increment of NAT equipment of the client A and are sent to the client B through an NAT traversal server; in the embodiment of the invention, N is 50, and Offset is 10;

step S223, client B binding Port_B3Sending a request to the 3479 Port of the STUN server to obtain a mapping Port_{B3_3479}Client B will Port_{B3_3478}N, Offset and Port increment Δ P_BSending the NAT traversal request to an NAT traversal server, forwarding the NAT traversal request to a client A by the NAT traversal server, and starting NAT traversal after the client A receives the NAT traversal request;

step S224, client A binding Port_A3Sending a request to the 3479 Port of the STUN server to obtain a mapping Port_{A3_3479}(ii) a Determining Port_{A3_3479}Whether or not to be greater than or equal to Port_{A2_3479}+Offset*ΔP_AIf the condition is not satisfied, the client A updates the local Port_A3Resending the request to the 3479 Port of the STUN server for other different values, and updating the mapping Port_{A3_3479}Repeating the above process until the mapping port satisfies the above condition; to reduce the impact of other processes or other clients on NAT device port assignment, client B's 1 st predicted port is calculated according to the following formula:

step S225, client A binding Port_A3Mapping N predicted ports within a Port prediction interval to client B sequentially_BpredSending a request in which Port_Bpred＝Port_{Bpred_1}+2*k*ΔP_BK is [0, N-1 ]]A natural number within the range;

step S226, when the client B receives the request of the client A at one of the monitoring ports, the NAT traversal is successful; and the client B sends a response to the client A, and the whole process is finished.

In step S2, when the NAT type of the client a is a local random symmetric type or a global random symmetric type, and the NAT type of the client B is an incremental symmetric type, as shown in fig. 7, the specific traversal process is as follows:

step S231, client B binds Port_B3Port, sending request to 3478 Port of STUN server to obtain mapping Port_{B3_3478}；

Step S232, client B binds N different local ports, and maps Port to client A' S forecast mapping Port in turn_ApredSending a request in which Port_ApredIs (P)_max,P_min) An integer randomly selected within the range; in the embodiment of the invention, N is 5;

step S233, client B binds Port_B3Port for sending request to 3479 Port of STUN server to obtain mapped Port_{B3_3479}(ii) a Client B will Port_{B3_3478}And Port_{B3_3479}Sending the NAT traversal request to an NAT traversal server, forwarding the NAT traversal request to a client A by the NAT traversal server, and starting NAT traversal after the client A receives the NAT traversal request;

step S234, client A binds local Port_A3Mapping each predicted Port within a Port prediction interval to client B sequentially_BpredSending a request message, wherein Port_Bpred＝Port_{B3_3478}+k*ΔP_BK is a positive integer, and Port_{B3_3478}<Port_Bpred<Port_{B3_3479}；

Step S235, when the client B receives the request of the client A, the NAT traversal is successful; the client B sends a response to the client A, and the traversing process is ended;

step S236, if the client A does not receive the response of the client B after time-out, the local Port is updated_A3If the value is different, returning to step S234 to retry, and retrying for Try times; in the present embodiment, Try is (P)_max-P_min) The smaller of the rounded value of/N and R.

In step S2, when the IP addresses of both clients are mapped_{X2_reflex}Similarly, and the host IP addresses of the two clients are different from the corresponding mapping IP addresses, it can be determined that the two clients are under the same NAT device, and the specific direct traversal process is as follows:

step S241, client A sends host address IP of client B_{B_host}:Port_B2Sending a request, IP_{B_host}:Port_B2Host address IP corresponding to client B_{X_host}:Port_X2The NAT passes through the server and is sent to the client A;

step S242, the client B receives the request packet sent by the client A and obtains the source address of the request packet; after obtaining the source address of the request packet, the client B sends a request to the address, and then the process goes to step S243; if the client B does not receive the request packet after timeout, go to step S244;

step S243, after receiving the request of the client B, the client A sends a response to the client B; if the client B receives the response, the direct traversing process is finished;

step S244, the client B sends the host address IP of the client A to the client B_{A_host}:Port_A2Sending a request, IP_{A_host}:Port_A2Host address IP corresponding to client A_{X_host}:Port_X2The NAT passes through the server and is sent to the client B;

step S245, the client A receives the request packet sent by the client B and obtains the source address of the request packet; after obtaining the source address of the request packet, the client a sends a request to the address, and then the process goes to step S246; if the client A does not receive the request packet after overtime, the direct NAT traversal fails, and the clients of the two parties can be regarded as different NAT devices and try again to perform NAT traversal by adopting a corresponding traversal strategy according to the NAT type;

step S246, after receiving the request of the client A, the client B sends a response to the client A; if the client A receives the response, the direct traversal process ends.

In step S2, when the NAT types of both clients are both port restricted cone type or firewall type, the method includes the steps of:

step S251, client A binding Port_A2Mapping an address IP to client B_{B2_reflex}:Port_{B2_3479}Sending a request in which Port_A2Port corresponding to client A_X2Port, IP_{B2_reflex}:Port_{B2_3479}Corresponding clientB mapping address IP_{X2_reflex}:Port_{X2_3479}The NAT passes through the server and is sent to the client A; because the client B does not actively send a request to the client A, the mapping does not exist in the NAT equipment of the client B, and the request can be intercepted by the NAT equipment of the client B; but leaves this mapping in client a's NAT device; after the client A sends the request, the port is monitored, and the request of the client B is waited.

Step S252, client B binding Port_B2Mapping an address IP to a client A_{A2_reflex}:Port_{A2_3479}Sending a request in which Port_B2Port corresponding to client B_X2Port, IP_{A2_reflex}:Port_{A2_3479}Mapping address IP corresponding to client A_{X2_reflex}:Port_{X2_3479}The NAT passes through the server and is sent to the client B; in step S251, the NAT device of the client a has already established a corresponding mapping, so the request of the client B can reach the client a; if the client A does not receive the request of the client B after overtime, the NAT traversal fails.

Step S253, after receiving the request of the client B, the client a sends a response to the client B, and since the NAT device of the client B has established the corresponding mapping in step S252, the client B can receive the response from the client a, and the process is ended after receiving the response; if the client B does not receive the response after time out, the NAT traversal fails.

In step S2, when the NAT type of the client a is other than the blocking type, the NAT type of the client B is the open type, and there is no NAT device, a "reverse connection" policy may be adopted, the client a first initiates a request, and the traversal process is as follows:

step S261, mapping address IP of client A to B_{B2_reflex}:Port_{B2_3479}Sending a request in which IP_{B2_reflex}:Port_{B2_3479}Mapping address IP corresponding to client B_{X2_reflex}:Port_{X2_3479}(ii) a The mapping is left in the NAT equipment of the client A, and the client A monitors the port after sending a request and waits for the request of the client B;

step S262, the client B receives the request packet sent by the client A and obtains the source address of the request packet; after obtaining the source address of the request packet, the client B sends a request to the address; if the client B does not receive the request packet after overtime, the NAT traversal fails;

step S263, after receiving the request of the client B, the client A sends a response to the client B; and after the client B receives the response, the traversing process is ended.

In step S2, when the NAT type of the client a is full cone type or IP restricted cone type, and the NAT type of the client B is other types except blocking type, the traversal process is as follows:

step S271, client A binding Port_A2Mapping IP address IP to B_{B2_reflex}Any 1 Port on, where Port sends the request_A2Port corresponding to client A_X2Port, IP_{B2_reflex}Mapping IP address IP corresponding to client B_{X2_reflex}The NAT passes through the server and is sent to the client A; the request may be intercepted by the NAT device of client B, but the mapping will be left in the NAT device of client a; after the client A sends a request, monitoring the port and waiting for the request of the client B;

step S272, mapping address IP from client B to client A_{A2_reflex}:Port_{A2_3479}Sending a request in which IP_{A2_reflex}:Port_{A2_3479}Mapping address IP corresponding to client A_{X2_reflex}:Port_{X2_3479}The NAT passes through the server and is sent to the client B; in step S271, the NAT device of the client a has already established a corresponding mapping, so the request of the client B can reach the client a; if the client A does not receive the request of the client B after overtime, the NAT traversal fails.

Step S273, after receiving the request of the client B, the client A sends a response to the client B; and after the client B receives the response, the traversing process is ended.

The above embodiments are preferred embodiments of the present invention, but the present invention is not limited to the above embodiments, and any other changes, modifications, substitutions, combinations, and simplifications which do not depart from the spirit and principle of the present invention should be construed as equivalents thereof, and all such changes, modifications, substitutions, combinations, and simplifications are intended to be included in the scope of the present invention.

Claims (9)

1. A NAT traversing method is applied to a system formed by a server and a client, wherein the server is deployed on a public network and comprises an NAT traversing server, a STUN server and a transit server, and the client comprises an NAT type detection module and an NAT traversing module, and is characterized by comprising the following steps:

s1, NAT type detection stage: the client side carries out NAT type detection by means of the STUN server and sends a detection result to the NAT traversal server, wherein the NAT type comprises a blocking type, an open type, a firewall type, a full cone type, an IP limiting cone type, a port limiting cone type, an incremental symmetry type, a local random symmetry type and a global random symmetry type; wherein, the STUN server includes 1 IP address and 2 ports, wherein, 2 ports are respectively port SP1 and port SP2, and the procedure of step S1 is as follows:

s101, the IP address of the host of the client X is IP_{X_host}Binding Port of client X_X1Port SP1 of access STUN server, obtaining Port from response message returned from STUN server_X1Mapped address IP of access port SP1_{X1_reflex}:Port_{X1_SP1}；

S102, if the overtime does not return, the system is in a blocking type, and the step S112 is entered;

s103, if IP_{X1_reflex}Equal to IP_{X_host}Checking whether the door is open or firewall type, namely executing step S104, otherwise executing step S106;

s104, Port of binding client X_X1A port SP1 of the STUN server is accessed again, the CHANGE port bit of the CHANGE-REQUEST attribute in the STUN command REQUEST message is set to be 1, and the STUN server is required to return a CHANGE port response through another port SP 2;

s105, if a change port response returned by the STUN server through a different port SP2 is received, the result shows that no NAT is converted, namely the STUN is an open type, otherwise, the result shows that the STUN is a firewall type, and the step S112 is executed;

s106, Port of binding client X_X1A port SP1 of the STUN server is accessed again, the CHANGE port bit of the CHANGE-REQUEST attribute in the STUN command REQUEST message is set to be 1, and the STUN server is required to return a CHANGE port response through another port SP 2;

s107, Port of binding client X_X1Port, Port SP2 of access STUN server, obtain Port_X1Mapped Port Port of Access Port SP2_{X1_SP2}；

S108, if Port_{X1_SP1}Equal to Port_{X1_SP2}Explaining that the NAT is a cone NAT, step S109 is executed to further distinguish which cone NAT is the NAT; otherwise, if the NAT is a symmetric NAT, step S110 is executed to further distinguish which symmetric NAT is;

s109, if a change port response returned by the STUN server through the port SP2 is received, the NAT is described as a full cone type or an IP restriction cone type, otherwise, the NAT is described as a port restriction cone type, and the step S112 is executed;

s110, binding different ports of client X_X2Port, Port SP1 of access STUN server, obtain Port_X2Mapped Port Port of Access Port SP1_{X2_SP1}(ii) a Port of binding client_X2Port, Port SP2 of access STUN server, obtain Port_X2Mapped address IP of access port SP2_{X2_reflex}:Port_{X2_SP2}；

S111, compare Port_{X1_SP1},Port_{X1_SP2},Port_{X2_SP1}And Port_{X2_SP2}A relationship of four;

if Port_{X1_SP2}-Port_{X1_SP1}＝Port_{X2_SP1}-Port_{X1_SP2}＝Port_{X2_SP2}-Port_{X2_SP1}If the NAT is in an increment symmetric type, the port increment is delta P; if not, the maximum value P of the four ports is calculated as a random symmetry type_maxAnd a minimum value P_minIf P is_maxAnd P_minIf the difference value of (A) is less than R, judging the local random symmetry type; if the R is larger than the R, the system is in a global random symmetry type;

S112、client side IP NAT type and host address_{X_host}:Port_X2And mapping address IP_{X2_reflex}:Port_{X2_SP2}Sending to NAT traversal server, wherein Port for open type, firewall type or cone type NAT_X2Get Port_X1Value of (2), IP_{X2_reflex}:Port_{X2_SP2}Mapping address IP_{X1_reflex}:Port_{X1_SP1}The detection process is ended;

s2, NAT traversal stage: after the NAT traversal server receives the NAT type detection results of the two communication client sides, the NAT traversal server determines different NAT traversal strategies, and then the NAT traversal strategies and the NAT type, the host address and the mapping address of the opposite client side, the port increment delta P of the increment symmetric NAT, the maximum value P of the measurement port of the local random symmetric type or the global random symmetric type NAT are used_maxAnd a minimum value P_minSending the data to the client of the other party; after receiving the NAT traversal strategy and the related information, the client sides of the two parties carry out NAT traversal according to the received NAT traversal strategy; and if the NAT traversal fails, forwarding by adopting a transit server.

2. The method according to claim 1, wherein in step S2, when the NAT type of the client a is a port restriction cone type or a firewall type, and the NAT type of the client B is an incremental symmetry type, the traversal process is as follows:

s201, client B binding Port_B3Port, sending request to Port SP1 of STUN server to obtain mapping Port_{B3_SP1}；

S202, binding Port by client B_B3Port, mapping address IP to client A_{A2_reflex}:Port_{A2_SP2}Sending a request in which IP_{A2_reflex}:Port_{A2_SP2}Mapping address IP corresponding to client A_{X2_reflex}:Port_{X2_SP2}The NAT passes through the server and is sent to the client B; client B then listens to Port_B3A port for waiting for a request message of a client A;

s203, binding Port by client B_B3Port, sends a request to port SP2 of the STUN server,get mapped Port_{B3_SP2}(ii) a Client B will Port_{B3_SP1}、Port_{B3_SP2}And the port increment delta P obtained in the NAT type detection process_BSending the data to an NAT traversal server; the NAT traversal server forwards the received message to the client A, and the client A starts NAT traversal after receiving the message;

s204, binding Port by client A_A2Port mapping each predicted Port in Port prediction interval to client B in sequence_BpredSending a request message, wherein Port_A2Port corresponding to client A_X2Port, Port_Bpred＝Port_{B3_SP1}+k*ΔP_BK is a positive integer, and Port_{B3_SP1}<Port_Bpred<Port_{B3_SP2}；

S205, when the client B is at the Port_B3When receiving the request of the client A, showing that the NAT traversal is successful; mapping Port of client B to client A_{A2_SP2}And sending a response to end the NAT traversal process.

3. The method according to claim 1, wherein in step S2, when the NAT type of the client a is a port-restricted cone type or a firewall type, and the NAT type of the client B is a local random symmetric type or a global random symmetric type, the traversal process is as follows:

s211, according to the success crossing probability P set by the system_sThe optimal parameters N and M are calculated by the following formula, where T is the total number of available ports, T65535 and 1024, and N and M are [1, T-]The calculated optimal parameters N and M can reduce the NAT traversal time;

s212, client B binds N different Port_B31、Port_B32……Port_B3NMapping Port to client A respectively_{A2_SP2}Sending requests, obtaining N different mapping ports in total, wherein the Port_{A2_SP2}Mapping Port corresponding to client A_{X2_SP2}The NAT passes through the server and is sent to the client B; then the client B monitors the N ports and waits for the request sent by the client A;

s213, client A binds local Port_A2Port, mapping Port Range to client B_{B2_SP2}-M/2+1,Port_{B2_SP2}+M/2]M ports within a Port send requests, where the Port sends the request_A2Port corresponding to client A_X2Port, Port_{B2_SP2}Mapping Port corresponding to client B_{X2_SP2}；

S214, when the client B receives the request of the client A at 1 monitoring port, the NAT traversal is successful; the client B maps the Port to the client A through the Port_{A2_SP2}And sending a response to end the NAT traversal process.

4. The method according to claim 1, wherein in step S2, when the NAT types of the clients of both parties are incremental symmetric types, the traversal process is as follows:

s221, client B binding Port_B3Sending a request to Port SP1 of the STUN server to obtain a mapping Port_{B3_SP1}；

S222, client B binding Port_B3Mapping N predicted ports within a Port prediction interval to client A sequentially_ApredSending a request; then monitoring the ports, and waiting for a request message sent by the client A; wherein, Port_Apred＝Port_{A2_SP2}+(k+Offset)*ΔP_AK is [1, N]Positive integers in the range, Offset is a natural number more than or equal to 0, and is used for avoiding an invalid port of the estimated NAT equipment of the client A; port (Port)_{A2_SP2}Mapping Port corresponding to client A_{X2_SP2}，ΔP_AThe parameters are port increment of NAT equipment of the client A and are sent to the client B through an NAT traversal server;

s223, client B binding Port_B3Sending a request to Port SP2 of the STUN server to obtain a mapping Port_{B3_SP2}Client B will Port_{B3_SP1}N, Offset and endIncrement of mouth Δ P_BSending the NAT traversal request to an NAT traversal server, forwarding the NAT traversal request to a client A by the NAT traversal server, and starting NAT traversal after the client A receives the NAT traversal request;

s224, client A binding Port_A3Sending a request to Port SP2 of the STUN server to obtain a mapping Port_{A3_SP2}(ii) a Determining Port_{A3_SP2}Whether or not to be greater than or equal to Port_{A2_SP2}+Offset*ΔP_AIf the condition is not satisfied, the client A updates the local Port_A3For other different values, the request is sent to Port SP2 of the STUN server again to update the mapping Port_{A3_SP2}Repeating the above process until the mapping port satisfies the above condition; to reduce the impact of other processes or other clients on NAT device port assignment, client B's 1 st predicted port is calculated according to the following formula:

s225, client A binding Port_A3Mapping N predicted ports within a Port prediction interval to client B sequentially_BpredSending a request in which Port_Bpred＝Port_{Bpred_1}+2*k*ΔP_BK is [0, N-1 ]]A natural number within the range;

s226, when the client B receives the request of the client A at 1 monitoring port, indicating that the NAT traversal is successful; and the client B sends a response to the client A to finish the NAT traversal process.

5. The method according to claim 1, wherein in step S2, when the NAT type of the client a is a local random symmetric type or a global random symmetric type, and the NAT type of the client B is an incremental symmetric type, the traversal process is as follows:

s231, client B binding Port_B3Port, sending request to Port SP1 of STUN server to obtain mapping Port_{B3_SP1}；

S232, binding N local different ports by the client B, and sequentially mapping the ports to the prediction mapping end of the client APort Port_ApredSending a request in which Port_ApredIs (P)_max,P_min) An integer randomly selected within the range;

s233, binding Port to client B_B3Port, sending request to Port SP2 of STUN server to obtain mapping Port_{B3_SP2}(ii) a Client B will Port_{B3_SP1}And Port_{B3_SP2}Sending the NAT traversal request to an NAT traversal server, forwarding the NAT traversal request to a client A by the NAT traversal server, and starting NAT traversal after the client A receives the NAT traversal request;

s234, client A binds local Port_A3Mapping each predicted Port within a Port prediction interval to client B sequentially_BpredSending a request message, wherein Port_Bpred＝Port_{B3_SP1}+k*ΔP_BK is a positive integer,. DELTA.P_BIs Port increment, and Port_{B3_SP1}<Port_Bpred<Port_{B3_SP2}；

S235, when the client B receives the request of the client A, showing that the NAT traversal is successful; the client B sends a response to the client A, and the NAT traversal process is ended;

s236, if the client A does not receive the response of the client B after time-out, the local Port is updated_A3And returning to the step S234 for retries for other different values, and retrying for Try times in total.

6. The method according to claim 1, wherein in step S2, when the mapping IP addresses of the two clients are IP addresses, the method further comprises_{X2_reflex}And the host IP addresses and the corresponding mapping IP addresses of the two client sides are different, the two client sides are judged to be under the same NAT equipment, and the traversing process is as follows:

s241, the host address IP from the client A to the client B_{B_host}:Port_B2Sending a request, IP_{B_host}:Port_B2Host address IP corresponding to client B_{X_host}:Port_X2The NAT passes through the server and is sent to the client A;

s242, the client B receives the request packet sent by the client A and obtains a source address of the request packet; after obtaining the source address of the request packet, the client B sends a request to the address, and then the process goes to step S243; if the client B does not receive the request packet after timeout, go to step S244;

s243, after receiving the request of the client B, the client A sends a response to the client B; if the client B receives the response, the direct traversing process is finished;

s244, the host address IP from the client B to the client A_{A_host}:Port_A2Sending a request, IP_{A_host}:Port_A2Host address IP corresponding to client A_{X_host}:Port_X2The NAT passes through the server and is sent to the client B;

s245, the client A receives the request packet sent by the client B and obtains the source address of the request packet; after obtaining the source address of the request packet, the client a sends a request to the address, and then the process goes to step S246; if the client A does not receive the request packet after overtime, the direct NAT traversal fails, and the clients of the two parties can be regarded as different NAT devices and try again to perform NAT traversal by adopting a corresponding traversal strategy according to the NAT type;

s246, after receiving the request of the client A, the client B sends a response to the client A; and if the client A receives the response, ending the direct NAT traversal process.

7. The method according to claim 1, wherein in step S2, when the NAT types of both clients are both port-restricted cone type or firewall type, the traversal process is as follows:

s251, client A binding Port_A2Mapping an address IP to client B_{B2_reflex}:Port_{B2_SP2}Sending a request in which Port_A2Port corresponding to client A_X2Port, IP_{B2_reflex}:Port_{B2_SP2}Mapping address IP corresponding to client B_{X2_reflex}:Port_{X2_SP2}The NAT passes through the server and is sent to the client A; after the client A sends a request, monitoring the port and waiting for the request of the client B;

s252, client B binding Port_B2Mapping to client AAddress IP_{A2_reflex}:Port_{A2_SP2}Sending a request in which Port_B2Port corresponding to client B_X2Port, IP_{A2_reflex}:Port_{A2_SP2}Mapping address IP corresponding to client A_{X2_reflex}:Port_{X2_SP2}The NAT passes through the server and is sent to the client B; if the client A does not receive the request of the client B after overtime, the NAT traversal fails;

s253, after receiving the request of the client B, the client A sends a response to the client B, and after receiving the response, the client B finishes the NAT traversal process; if the client B does not receive the response after time out, the NAT traversal fails.

8. The method according to claim 1, wherein in step S2, when the NAT type of the client a is other than blocking type, the client B is open type, and there is no NAT device, the traversal process is as follows:

s261, mapping address IP of client A to B_{B2_reflex}:Port_{B2_SP2}Sending a request in which IP_{B2_reflex}:Port_{B2_SP2}Mapping address IP corresponding to client B_{X2_reflex}:Port_{X2_SP2}(ii) a After the client A sends a request, monitoring the port and waiting for the request of the client B;

s262, the client B receives the request packet sent by the client A and obtains the source address of the request packet; after obtaining the source address of the request packet, the client B sends a request to the address; if the client B does not receive the request packet after overtime, the NAT traversal fails;

s263, after receiving the request of the client B, the client A sends a response to the client B; and after receiving the response, the client B ends the NAT traversal process.

9. The method according to claim 1, wherein in step S2, when the NAT type of the client a is full cone type or IP restricted cone type, and the NAT type of the client B is other types except blocking type, the traversal process is as follows:

s271, binding end of client APort Port_A2Mapping IP address IP to B_{B2_reflex}Any 1 Port on, where Port sends the request_A2Port corresponding to client A_X2Port, IP_{B2_reflex}Mapping IP address IP corresponding to client B_{X2_reflex}(ii) a After the client A sends a request, monitoring the port and waiting for the request of the client B;

s272, mapping address IP from client B to client A_{A2_reflex}:Port_{A2_SP2}Sending a request in which IP_{A2_reflex}:Port_{A2_SP2}Mapping address IP corresponding to client A_{X2_reflex}:Port_{X2_SP2}(ii) a If the client A does not receive the request of the client B after overtime, the NAT traversal fails;

s273, after receiving the request of the client B, the client A sends a response to the client B; and after the client B receives the response, the traversing process is ended.

Images