JP2008219460A - Communication interface switching method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a communication system by which application communication can be continued even if network interfaces are switched by such a reason as that a system is moved in a home network. <P>SOLUTION: In this system, a system suitable for an environment of such a home network that information about equipment connected to the home network is obtained and a technique optimal to situations such as the case of switching to then interface of the same subnet and the case that a communication partner has a similar system is adopted. This communication interface switching system is characterized by being provided with a means for performing advertisement by a Gratuitous ARP; a means for continuing communication by performing the IPinIP encapsulation of an IP header and a means for performing the setting of a fixed transmitting address, address conversion and/or a packet relay by the home network HA. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a communication interface switching method. More specifically, the present invention relates to a communication interface switching method for continuing application level communication.

  In recent years, home networks have attracted attention, and products that connect various home appliances such as televisions, video recorders, air conditioners, refrigerators, etc. to the network and perform mutual cooperation have been put into practical use. Standardization of methods for mutual recognition and data transfer between home information appliances such as DLNA is also being carried out. Thus, as home networks become widespread, the diversity of home networks becomes an issue to be considered. In home networks, a plurality of networks such as Ethernet (registered trademark), wireless LAN, and power line LAN are introduced, and it is assumed that information home appliances have corresponding communication interfaces. In that case, the network to be connected during communication may be changed depending on the usage status of the information home appliance. For example, when a portable TV moves to another room while the portable TV and the video server are communicating via a wireless LAN, the wireless communication condition deteriorates and the Ethernet (registered trademark) is connected. The situation is such that communication is continued. In such a situation, the communication interface to be used is changed, and therefore the IP address on the terminal side is changed, so that the application level communication up to that point must be terminated once. However, considering user convenience, an optimal communication interface is selected independently of application communication, and a function is required to continue communication even when the interface is changed.

The function to continue communication even if the IP address of the terminal is changed is Mobile to support the movement of the terminal in the mobile network
This is realized by IP (Non-Patent Document 1). However, Mobile IP assumes a wide area network environment, and HA is necessary for each subnet, and when moving, registration to HA is always required. There is.

  On the other hand, as a method for switching devices connected to the same subnet, a method is known in which another device continues communication using the IP address of the original device using gratuitous ARP (Non-patent Document 2). . If this method is used, for example, an interface can be switched between Ethernet (registered trademark) and a wireless LAN in the same subnet. However, this method alone cannot cope with interface switching across subnets, such as switching from direct communication using wireless LAN ad hoc mode to communication using Ethernet (registered trademark).

C. Perkins, Ed., “IP Mobility Support for IPv4,” RFC 3344, Aug. 2002. A. Bhide, et al., “A Highly Available Network File Server,” Proc. Of 1991 Winter USENIX Conference, pp.199-205, 1991.

  An object of the present invention is to provide a communication method that can continue application communication even when a network interface is switched due to a movement of the system in a home network.

  Against this background, this system proposes a new procedure that enables information home appliances with multiple communication interfaces to switch interfaces due to movement or changes in radio wave conditions while maintaining application-level communication. To do. Specifically, it is possible to obtain information related to devices connected to the home network, use a method that best suits the situation, such as when switching to an interface on the same subnet or when the communication partner is a similar system. A method suitable for the network environment will be adopted.

Therefore, the present invention is a communication interface switching method, and a gratuitous
Communication comprising means for advertising by ARP, means for continuing IPinIP encapsulation of the IP header, and means for setting a fixed transmission address and performing address translation and / or packet relay by the home network HA Provides an interface switching method.
In addition, the present invention is a communication interface switching method, and when the switching destination interface of the own system belongs to the same subnet, the switching is performed with respect to the switching source IP address by Gratuitous ARP when switching the interface. When switching the interface when the MAC address of the previous interface is supported and the interface to which the system is switched is connected to a different subnet and the partner system supports the proposed method A means for sending a switch request to a communication partner, and for both systems to continue communication by performing IPinIP encapsulation of the IP header using the IP address of the new interface and the IP header using the IP address of the original interface , The interface to which your system is switched When connected to another subnet and the partner system does not support the proposed method, the home network HA in the home network has a means for setting a fixed transmission address and performing address translation / packet relay Provides a communication interface switching method.
Furthermore, the present invention is the communication interface switching method described above, wherein communication is performed between the own system and the home network HA, and between the home network HA and the partner system, and between the own system and the home network HA. In the meantime, IPinIP encapsulation is used, while the home network HA performs NAT to provide a communication interface switching method in which the address of the own system is always kept constant and the address is changed.
Moreover, this invention provides the apparatus provided with the said communication interface switching system.

  According to the present invention, it is possible to provide a communication method capable of continuing application communication even when the network interface is switched due to the movement of the system in the home network.

  The communication interface switching method of the present invention will be described below with reference to the drawings.

  As shown in Fig. 1, this system is a network in which many information appliances with multiple interfaces such as Ethernet (registered trademark), wireless LAN, and power line LAN are connected, and there are multiple subnets and communication paths in the LAN. In the environment, even if each node connected to the network switches the active interface during communication, the object is to continue application level communication.

Mobile, which is widely used to support node movement in mobile networks, in order to hide IP address changes associated with switching of communication interfaces from applications
Based on IP. However, Mobile IP is designed to be used in a wide area network environment. HA is required for each subnet, and it is always necessary to register with HA when moving. Large overhead. Therefore, this method employs the following method.

  (1) Even in a home network with multiple subnets, install one home network HA (hereinafter referred to as home network HA) that supports this method.

  (2) If multiple interfaces of a node are all connected to the same subnet, apply the method using Gratuitous ARP when switching interfaces.

  (3) If the node needs to be switched to an interface connected to a different subnet, and the communicating node supports this method, both parties cooperate and continue application-level communication when the interface is switched.

  (4) If the above two conditions are not met, that is, if the node needs to be switched to an interface connected to a different subnet, and the communicating node does not support this method, it is assumed that the interface will be switched. As for the communication to be performed, communication is always performed via the home network HA from the start of communication, and the node and the HA perform interface switching in cooperation.

  In the home network, it is possible to obtain information on other nodes connected to the LAN when each node is connected to the network by UPnP or the like. In this method, each node can acquire the following information from each other. The information includes information on whether or not the proposed method is supported, and information on the type and address of the implemented interface.

  Taking advantage of the characteristics of the home network that information on other devices connected in the network can be acquired, the communication method is switched as follows according to the function of the communication partner node.

If the correspondent node does not support the proposed method and there is a possibility of switching the communication interface to an interface belonging to a different subnet during communication, bidirectional tunneling in Mobile IP using the home network HA [G. Montenegro, Ed ., “Reverse Tunneling
for Mobile IP, ”RFC 2344, May 1998] and NAT are used for communication. Communication is performed between the own node and the home network HA, and between the home network HA and the partner node. IPinIP encapsulation is used between the own node and the home network HA, and the home network HA performs NAT so that the address of the own node is always kept constant and the address is changed.

  In cases other than the above, that is, when there is no possibility of switching the communication interface to an interface belonging to a different subnet even when communicating with a node that supports the proposed method or does not support the proposed method, the home network Communication using HA is not performed, but normal communication is performed until the communication interface is switched.

  When communication is performed without using the home network HA and the communication interface is switched to an interface belonging to the same subnet, a method using Gratuitous ARP is employed to continue communication. In this method, the IP address of the switching source interface is advertised so as to correspond to the MAC address of the switching destination interface, and communication is continued.

When communication is performed without using home network HA and the communication interface is switched to an interface belonging to a different subnet, an interface switch request is sent directly to the communication partner when the interface is switched.
IP Route Optimization [C. Perkins and D. Johnson, “Route Optimization for Mobile IP,” J.
Cluster Computing, Vo.1, No.2, pp.161-171, Jun. 1998]. In this method, an interface switch request is sent to the communication partner when the interface is switched, and both systems perform IPinIP encapsulation of the IP header using the IP address of the new interface and the IP header using the IP address of the original interface. Go and continue communication.

  In the above description, methods using Gratuitous ARP, Mobile IP (its route optimization), bidirectional tunneling, NAT, and IPinIP encapsulation are used in each situation, but instead of these methods, It will be apparent to those skilled in the art that various approaches corresponding to these approaches may be used.

Sequence example 1. Communication method using home network HA
FIG. 3 shows a sequence example when the communication interface is switched from Ethernet (registered trademark) to a wireless LAN while the MN is communicating with a general CN on the Internet as shown in FIG. In this example, CN is a general node on the Internet and does not support the proposed method. A MN with an Ethernet (registered trademark) and wireless LAN interface to which IPs of different subnets are assigned can communicate with the CN even using the wireless LAN interface, but here it starts communication with Ethernet (registered trademark). Suppose that The MN starts communication using the home network HA because the CN does not support the proposed method and the communication interface may be switched to an interface belonging to a different subnet during communication.

First, the MN interfaces with the home network HA
Send a switching registration request message. This message includes information that can identify the application level communication. That is, the IP address of the Ethernet (registered trademark) interface of the MN, the IP address of the CN, and the port numbers of the MN and CN. Upon receiving this message, the home network HA secures a new translated IP address and port number for the application level communication identification information included in the message. The secured conversion IP address and port number are associated with application level communication identification information included in the message and recorded. Furthermore, HA is an interface
A switching registration response message is sent to the MN.

Upon receiving this response message, the MN starts communication through the Ethernet (registered trademark) interface. In this case, the IP datagram is IPinIP encapsulated and transmitted to the HA. As shown in 1 of Figure 3, for the outer IP header of this IP datagram, the destination IP address is the IP address of the home network HA, the source IP address is the IP address of the MN's Ethernet (registered trademark) interface, and the inner IP address As for the IP header, the destination IP address is the IP address of the CN, and the source IP address is the IP address of the Ethernet (registered trademark) interface of the MN. When the home network HA receives the encapsulated datagram, it removes the outer IP header. Furthermore, from the inner IP header and port number interface
Search the application level association information created when receiving the switching registration request message, and if the corresponding information is found, convert it to the registered source IP address and port number in the IP header and forward to CN (Fig. 3 of 2). This conversion process corresponds to NAPT operation by the home network HA.
Datagrams sent from the CN are delivered via the reverse route. The IP header of the datagram between CN and home network HA is as shown in 3 of FIG. When the home network HA receives the IP datagram, it searches the application-level association information registered from the IP header and port number of the datagram, and if the corresponding information is found, it matches the source IP address and port number. It is converted into the one and IPinIP encapsulated as shown in 4 of FIG.

Here, it is assumed that the MN detects disconnection of the Ethernet (registered trademark) interface. Then, the MN uses the wireless LAN interface to interface with the home network HA.
Send a switching notification message. This message includes information that can identify the application level communication. That is, it includes the IP address of the MN's Ethernet (registered trademark) interface, the IP address of the CN, the port numbers of the MN and CN, and the IP address of the wireless LAN interface. When the home network HA receives this message, it searches the registered application level association information, and updates the information if there is a corresponding one. In addition to MN interface
Returns a switching notification acknowledgment message. When the MN receives this acknowledged message, the MN continues communication through the wireless LAN interface. Note that while waiting for this acknowledgedment message, the MN buffers transmission datagrams in this application level communication. interface
After switching notification, the IP header of the datagram between the MN and the home network HA is as shown in 5 of FIG. 3, 9 between the home network HA and CN, 7 between the CN and home network HA, and the home network HA- It becomes like 8 between MN.

Above is Mobile
It is a method similar to bidirectional tunneling in IP. However, in this method, unlike Mobile IP, the MN does not use a home address. The CN communicates with the address (and port) generated by the HA.

2. Communication without using home network HA: When switching to an interface belonging to the same subnet In the topology shown in Fig. 4, the MN with two interfaces, Ethernet (registered trademark) and wireless LAN, is an Ethernet (registered trademark) interface. FIG. 5 shows a sequence example when the communication interface is switched to the wireless LAN while the communication is started using and the communication is continued. In the topology of FIG. 4, the wireless LAN access point is connected to the same Ethernet (registered trademark) as the Ethernet (registered trademark) to which the MN's Ethernet (registered trademark) interface is connected. The CN is connected to the Ethernet (registered trademark) to which the access point is connected, but does not have a wireless LAN interface.

The MN performs normal communication until a communication interface switching process occurs. When the MN detects that the Ethernet (registered trademark) interface is disconnected during communication and finds that another interface connected to the same subnet, that is, a wireless LAN interface is available, the following processing is performed. First, the Ethernet (registered trademark) interface is explicitly brought down, and the IP address previously assigned to the Ethernet (registered trademark) interface is assigned to the wireless LAN interface (thereby assigning two IP addresses to the wireless LAN interface) Will be.) At the same time, gratuitous from the wireless LAN interface
Send an ARP request message. The parameters of this Gratuitous ARP message are as follows.

Source IP address: Ethernet (registered trademark) IP address Source MAC address: Wireless LAN MAC address Target IP address: Ethernet (registered trademark) IP address.

CN is this ARP
When the request is received, the ARP cache is updated so that the IP address of the MN's Ethernet (registered trademark) interface is mapped to the MAC address of the MN's wireless LAN by the ARP function. This allows the CN to send an IP datagram with the destination MAC address as the MAC address of the MN's wireless LAN interface and the destination IP address as the IP address of the MN's Ethernet (registered trademark) interface. Can continue.

3. Communication without using home network HA: When switching to an interface belonging to the same subnet As shown in FIG. 6, both the MN and CN have two interfaces, Ethernet (registered trademark) and wireless LAN. Furthermore, it is assumed that the MN and the CN are set so that they can communicate with each other in the wireless LAN ad hoc mode. The IP address used in the ad hoc mode has a different subnet from Ethernet (registered trademark). FIG. 7 shows a sequence example when both the MN and the CN start communication using the Ethernet (registered trademark) interface in such a topology, and the communication interface is switched to the ad hoc mode wireless LAN while continuing the communication.

The MN performs normal communication until a communication interface switching process occurs. If the MN detects that the Ethernet (registered trademark) interface is disconnected during communication and finds that another interface connected to a different subnet, that is, a wireless LAN interface in ad hoc mode, can be used, the following processing is performed. Do. First, the MN is interfaced from the wireless LAN interface.
Send a switching request message. The request message includes information that can identify the application level communication. In other words, this information includes the IP address of the Ethernet (registered trademark) interface of the MN, the IP address of the Ethernet (registered trademark) interface of the CN, and the port numbers of the MN and CN. When the CN receives this message, it registers the application level identification information contained in the message and records that this communication needs to be continued via the wireless LAN interface. Furthermore, interface to MN
Send a switching response message. When the MN receives this message, it registers the application level identification information included in the message and records that this communication needs to be continued via the wireless LAN interface. During the response message waiting state, the MN buffers transmission datagrams in this application level communication. interface
When the exchange of switching messages is completed, the MN and CN encapsulate all IP datagrams for the target application level communication, and resume communication. The outer IP headers of these IP datagrams contain the wireless LAN IP addresses of the MN and CN. The encapsulated IP datagram is transferred through a wireless LAN in ad hoc mode.

The above method is similar to the route optimization of Mobile IP. However, this method does not have HA support for Mobile IP route optimization. Mobile
Unlike IP route optimization, it is managed for each session.

4. Flowchart In this section, detailed processing of the MN in this method is shown as a flowchart.

  FIG. 8 shows processing when the MN is connected to the home network. As described above, the MN obtains information regarding other nodes connected to the home network. The information includes the logical name of the MN, information on whether or not the proposed method is supported, all the interfaces of the node, the IP address and subnet mask assigned to it, and the ad hoc radio on which the terminal is independent This is information on whether or not connected to the LAN segment.

FIG. 9 shows processing when the MN starts application-level communication. First, the MN determines whether to apply the interface switching method to the communication to be started. In this method, it is assumed that the determination whether or not to perform communication using this interface switching method is performed for each application by the user program. If the interface switching method is applied, the MN checks whether or not to use the home network HA. When home network HA is used, MN is interface
Send a switching registration request message to the home network and receive the response. Here, the MN records that the interface switching method is applied to this communication.

  FIG. 10 shows processing when the MN transmits an IP datagram. First, the MN checks whether or not to apply the interface switching method to the IP datagram to be transmitted. If not applied, the MN does nothing and sends an IP datagram. If the interface switching method is applied, the MN next checks whether or not the communication uses the home network HA. If this application-level communication uses the home network HA, the MN uses the outer IP header of the IP datagram as the source IP address as the IP address of the MN and the destination IP address as the IP address of the home network HA. To do.

  When the home network HA is not used, it is checked whether the new interface of the MN belongs to the same subnet as the current communication interface. If it belongs to the same subnet, the MN sends the IP datagram without doing anything. If they belong to the same subnet, processing at the ARP level is executed even if the interface is switched, so IP level processing is not necessary.

  Otherwise, the MN negotiates directly with the CN and performs interface switching. If the switch has not yet been performed, the IP datagram is sent as is. When switching is performed, the IP datagram is IPinIP encapsulated and transmitted based on the application level communication information registered at the time of negotiation.

FIG. 11 shows processing when interface switching is executed. The interface switching process is started when the interface is disconnected, the communication quality deteriorates, or an interface with a higher priority becomes available. When the interface switching process is started, the MN executes the interface switching process for all the application level communications being executed. If communication is performed using the home network HA, the MN
Send a switching notification request message to the home network HA. In the case of communication not using the home network HA, the MN checks whether the new interface belongs to the same network, and if so, performs the processing described above. That is, the interface that has been used so far is taken down and its IP address is assigned to the new interface. Further parameter gratuitous
Send ARP. If the new interface belongs to a different subnet, the MN will directly negotiate with the CN
A switching request message is transmitted to the CN, and an interface switching response message is received. Here, how to select the IP address of CN will be described. Since MN obtains information about CN by UPnP, the IP address of CN is an independent ad.
It can be determined whether the wireless LAN segment is in the hoc mode.

If the CN's IP address used before is an independent ad
If the wireless LAN segment is in the hoc mode, the MN specifies the IP address of the CN connected to the backbone of the home network, in addition to the IP address used so far, in the interface switching request message. If the CN IP address used so far is connected to the backbone of the home network, communication will continue using the same CN IP address after interface switching. Here, if the CN has an IP address that belongs to the same subnet as the IP address of the new interface of the MN and can be used, communication may be continued using that IP address.

A home network assumed to use the communication system of the present invention. The figure which shows the topology (communication using HA) of a sequence example. The figure which shows the example of a sequence (communication using HA). The figure which shows the topology (communication which does not use HA: switching to the same subnet) of the sequence example. The figure which shows the example of a sequence (communication which does not use HA: Switching to the same subnet). The figure which shows the topology (communication which does not use HA: switching to a different subnet) of the sequence example. The figure which shows the example of a sequence (communication which does not use HA: Switching to a different subnet). The figure which shows a process when MN is connected to a home network. The figure which shows a process when MN starts application level communication. The figure which shows the process in case MN transmits IP datagram. The figure which shows a process when interface switching is performed.

Claims (4)

A communication interface switching method,
A means of advertising via Gratuitous ARP;
Means to continue communication by performing IPinIP encapsulation of the IP header;
Means for performing fixed transmission address setting and address translation and / or packet relay by the home network HA;
Communication interface switching method with A communication interface switching method,
When the interface to which the system is switched belongs to the same subnet, means to advertise so that the MAC address of the interface to be switched corresponds to the IP address to be switched by Gratuitous ARP when switching the interface When,
When the interface to which the system is switched is connected to a different subnet and the partner system supports the proposed method, a switching request is sent to the communication partner when the interface is switched, and both systems have new interfaces. Means to continue communication by performing IPinIP encapsulation of the IP header using the IP address of the original interface and the IP header using the IP address of the original interface;
If the interface to which your system is switched is connected to a different subnet, and the remote system does not support the proposed method, the home network HA in the home network will set a fixed transmission address and address translation / packet Means for relaying,
Communication interface switching method with The communication interface switching method according to claim 1 or 2,
Communication is performed between the own system and the home network HA, and between the home network HA and the partner system. The IP network uses IPinIP encapsulation between the own system and the home network HA. A communication interface switching method that performs NAT and always shows the address of the local system as well as changing the address.   An apparatus comprising the communication interface switching method according to any one of claims 1 to 3.