The present invention relates to a method for switching network traffic between network interfaces of a networked device. In particular, the present invention relates to seamless switching of network traffic between interfaces. The present invention also relates to a networked device for performing the method. The present invention further relates to a computer program for performing the method.

In the future, many mobile or stationary devices which are connected to the Internet will be expected to have multiple network interfaces. This may, e.g., be a wired and a wireless network interface. Examples of such devices are Personal Digital Assistants (PDA's), notebooks, laptop computers, etc. In a home network environment it is very possible that a wired as well as a wireless interface are connected to the same link. For example, a PDA which connects wirelessly to the home server may also connect to the home server via its wired interface once it is put in the cradle, and the cradle and the wireless access point are connected, e.g. via a hub. A similar scenario may exist for a notebook or a laptop computer. One network interface may be preferable with respect to one or more other network interfaces, e.g. due to speed, costs, reliability, etc. It may, thus, be desirable to redirect communication from one network interface of a device to another if these network interfaces are on the same link. Typically, such a redirection of communication is performed by closing down the connection on the existing interface and setting up a new connection on the other interface. This is a cumbersome and time consuming operation since the existing communication path must be stopped and set up again. Furthermore, there is a risk that messages will be lost during the closing down of the connection. Thus, it is desirable to provide a method for redirecting the communication seamlessly, i.e. without having to close down the existing connection. The article Mobile IP2, 12th IEEE International Symposium on Personal, Indoor and Mobile Radio Communications, PIMRC 2001, proceedings, vol. 1, pp. A-88-92, discloses a method for high-speed handoff and seamless roaming between different networks of a mobile device. Multiple connections between a mobile device and a Home Agent are possible via various Foreign Agents, and the Home Agent maintains a table consisting of multiple Foreign Agent entries. The Home Agent will tunnel a given data packet to an appropriate Foreign Agent based on the user profile. One disadvantage of the method disclosed in the article is that it requires the presence of a Home Agent. Thus, in case the Home Agent does not exist or if it for some reason is unreachable, the method described in the article does not work. Furthermore, packets will be routed over the Home Agent, and if this is located far away, it will not be efficient, even if source and destination are on the same link. Thereby communication on the local area network is made inefficient. Another disadvantage is that the method requires modifications to existing protocols in order to function, e.g. by adding enhancements to the Address Resolution Protocol (ARP) in order to control 'physical' routing of Internet Protocol (IP) packets for multiple connections. This makes the method relatively difficult to implement.

It is, thus, an object of the present invention to provide a method for redirecting communication between network interfaces of a networked device which does not require the presence of a Home Agent. It is a further object of the present invention to provide a method for redirecting communication between network interfaces of a device which can be implemented employing existing protocols without requiring modifications to these. It is an even further object of the present invention to provide a method for redirecting communication between network interfaces of a networked device in a seamless manner, i.e. without having to close down the connection on the existing interface and setting up a new one. It is an even further object of the present invention to provide a method for redirecting communication between network interfaces of a networked device which improves the efficiency of the communication on the local area network. According to a first aspect of the present invention, the above and other objects are fulfilled by providing a method for redirecting communication from a first network interface of a networked device to a second network interface of said networked device by means of switching an internet protocol (IP) address from said first network interface to said second network interface, said first and second network interfaces being on the same link, the method comprising the steps of: assigning and/or adding to said second network interface an IP address, IPO, associated with said first network interface, adding, in a route table of the device, the entry of the second network interface related to the IPO address, - deleting, from the route table of the device, the entry/entries of said first network interface related to the IPO address, and updating one or more traffic management tables located in one or more remote hosts. According to a second aspect of the present invention, the above and other objects are fulfilled by providing a networked device comprising at least a first and a second network interface, the device further comprising: means for assigning and/or adding to said second network interface an IP address, IPO, associated with said first network interface, means for adding, in a route table of the device, the entry of the second network interface related to the IPO address, means for deleting, from the route table of the device, the entry/entries of said first network interface related to the IPO address, and means for updating one or more traffic management tables located in one or more remote hosts. The effect of this procedure is that any communication which initially went via the first network interface is now going via the second network interface. Furthermore, the switching of the traffic has been performed in such a way that it has not been necessary to close the original connection and open a new one. Since all the steps of the method according to the present invention are performed locally on the device the requirement of a Home Agent is avoided. Thereby the transition of the traffic is performed transparently, i.e. without requiring dedicated actions with remote hosts. Furthermore, the method can be easily implemented on existing devices since this only requires local modifications, i.e. modifications to the device itself. These modifications may even be performed by merely adding software, i.e. no physical modifications to the device are necessary. Furthermore, also due to the local operations, the method can be implemented using existing protocols, existing route tables, existing traffic management tables, etc., without requiring modifications to these. This is a great advantage since the implementation of the method is thereby substantially facilitated as compared to methods needing modifications to existing protocols etc. in order to function correctly. When in the present context the term 'being on the same link' is used, this should be interpreted as the two or more network interfaces in question being able to reach substantially the same set of network nodes. Thus, the ingoing/outgoing communication of the device does not depend on which of the network interfaces is chosen. However, there will typically still be differences between the interfaces with respect to, e.g., speed, cost, stability, etc., and these differences may encourage a user to prefer one interface to other interfaces if several are available. The first network interface is the interface which the device is initially using for communication, and the IP address, IPO, is accordingly the IP address which is initially used for communication. The steps of the method are preferably performed in the stated order. Thus, because the IP address, IPO, is assigned/added to the second network interface before adding the entry of the second network interface related to the IPO address in the route table of the device, which is in turn performed before deleting the entry/entries of the first network interface, a substantially seamless transition between the interfaces is obtained. That is, it is not necessary to close the original connection and set up a new one in order to switch traffic from one network interface to another. In other words, all transmission control protocol (TCP) connections stay open during the switching and no user datagram protocol (UDP) packets are lost. The updating step may be performed by broadcasting unsolicited address resolution protocol (ARP) messages. In this case the one or more traffic management tables being updated are ARP tables. ARP being a standard protocol, it is very advantageous that ARP messages can be used without requiring any modifications. Furthermore, since ARP messages do not cross links, all interfaces reached by an ARP message are, by definition, on the same link, and it is therefore very suitable to use ARP messages for this purpose. The remote hosts in which the traffic management tables are located may, e.g., be any suitable computing device having at least one network interface. Thus, it could, e.g., be a PDA, a notebook, a laptop computer, a desktop computer, or any other suitable kind of computing device. The method may further comprise the step of switching off the first network interface and/or removing the IPO address from the first network interface. This has the advantage that power is saved. Furthermore, possible conflicts and/or unexpected behaviors are avoided in case remote hosts keep on using the first network interface, e.g. because they did not receive the ARP messages. However, in some applications it is not necessary to switch off the first network interface and/or remove the IPO address from it. The method may further comprise the step of determining whether two or more network interfaces are on the same link. In this case this determining step should be performed prior to performing the assigning/adding step in order to ensure that the two network interfaces are indeed on the same link before starting the switching procedure. The determining step may advantageously be performed by sending a message from the first network interface, receiving a message on the second network interface and correlating the sent and the received messages. Based on the outcome of the correlating step it is determined whether or not the two network interfaces are on the same link. The correlation may, e.g., involve checking whether or not the message has crossed a level-3 router. However, the determining step may be performed in any suitable manner. In case it is determined that two or more network interfaces are on the same link, the method may further comprise the step of selecting a desired one of said network interfaces based on one or more predetermined criteria. In this case the remaining steps are only performed in case a network interface which is different from the currently used network interface is selected as the desired network interface. Thus, in case the currently used network interface is the most advantageous of the available network interfaces no switching is performed. On the other hand, in case one of the other available interfaces is more advantageous than the one currently used, the traffic should be switched to this one. Additionally, switching may also be performed in case it is determined that the currently used network interface is no longer on the link. In this case the traffic should be switched to another network interface which is still on the link. The method may further comprise the steps of: determining that the first and the second network interfaces are no longer on the same link, assigning and/or adding the IPO address to the first network interface, - adding, in the route table of the device, the entry of the first network interface related to the IPO address, deleting, from the route table of the device, the entry/entries of the second network interface related to the IPO address, and updating one or more traffic management tables located in one or more remote hosts. According to this embodiment, in case the network topology changes in such way that the two interfaces are no longer on the same link, it is possible to perform a rollback and move traffic back to the first network interface again. This is done exactly as the switching of traffic from the first network interface to the second network interface as described above, i.e. in a seamless and transparent manner. Topology changes may e.g. be caused by a user removing the device from a cradle, thereby disrupting the connection to the wired network. As a consequence the wired and the wireless interfaces will no longer be on the same network link. The method of the present invention may advantageously be performed by a computer program. The computer program may, e.g., be running in a general purpose computer, be stored on a carrier, and/or be transferred over a network, etc. These and other aspects of the invention will be apparent from and elucidated with reference to the embodiments described hereinafter.

Fig. 1 shows the initial situation at the device side, Fig. 2 shows the initial situation at the remote host side, Figs. 3-7 illustrate various steps of one embodiment of the method of the present invention, and Fig. 8 shows a networked device according to an embodiment of the present invention.

Fig. 1 is an illustration of an initial situation for a device having two network interfaces, If_A 1 and If_B 2, each being associated with an IP address, IP_A and IP_B, respectively. The two interfaces, If_A 1 and If_B 2, are both on the same network link. According to the route table 3 of the device, in this initial situation, all the traffic to hosts in the IP_A subnet is sent via If_A 1, and all the traffic to hosts in the IP_B subnet is sent via If _B 2. Fig. 2 also illustrates an initial situation, but seen from the point of view of a remote host. The remote host has IP address IP_C and network interface If_C 4. The information in the ARP table 5 of the remote host corresponds to the situation in the device as shown in Fig. 1 and described above. Thus, in the ARP table 5 of the remote host IP address IP_A is associated with MAC address If_A and IP address IP_B is associated with MAC address If_B. Accordingly, all traffic to IP address IP_A is sent to MAC address If_A, and all traffic to IP address P B is sent to MAC address IfJB. Figs. 3-7 illustrate various steps of an embodiment of the method of the invention. The figures all show the situation in the device as well as in the remote host with the changes taking place in connection with the relevant step being highlighted. In the illustrated embodiment network traffic is switched from interface If B 2 to interface If_A 1. In Fig. 3 the IP address IP_B is added/assigned to interface If_A 1. Thus, in Fig. 3 IP_B is assigned to both If_A 1 and If_B 2. Next, as illustrated in Fig. 4, an entry 6 of IP B subnet associated with If_A 1 is entered in the route table 3 of the device. Thereby IP_B subnet is associated with both If A 1 and If_B 2. When sending a message the route table 3 picks the first entry which matches the destination address. Therefore, in the situation illustrated in Fig. 4, the outgoing traffic to hosts in the IP_B subnet is still sent via If_B 2 because this entry is above the new entry associating the IP_B subnet with If_A 1. In the step illustrated in Fig. 5 the route table entry 7 of IP B associated with If B 2 is removed. Thus, IP_B subnet is only associated with If_A 1, and consequently all outgoing traffic to hosts in the IP_B subnet is now sent via If A 1. However, since no changes have yet been performed to the ARP table 5 of the remote host, the remote host will continue to send traffic to IP_B on If B 2 in accordance with the information in the ARP table 5. In the step illustrated in Fig. 6 the device sends unsolicited ARP replies 8 containing IP address IP_B. The replies 8 are sent from interface If_A 1. Upon receipt of the replies 8 the remote host updates the ARP table 5 in such a way that IP address IP_B is now associated with interface If_A 1. This is indicated at entry 9. As a consequence all incoming traffic to IP_B is now sent to If_A 1 and no longer to If_B 2. In Fig. 7 the optional step of removing IP_B from If_B 2 or closing If_B 2 is illustrated. Fig. 8 shows a networked device 10 according to an embodiment of the present invention. The networked device 10 comprises two network interfaces, If_A 1 and If_B 2, each being associated with an IP address, IP_A and IP_B, respectively. The two interfaces, If_A 1 and If_B 2, are both on the same network link 11. The networked device 10 furthermore comprises a processor unit 12, for controlling the network interfaces 1 and 2, and thereto connected to these interfaces via connections 13 and 14. The processor unit 12 is furthermore coupled via a connection 15 to a memory 16. Typically, the connections 13, 14 and 15 are implemented as one or more data and/or control busses. The memory 16 comprises a route table 3 of the networked device 10 and a program memory 17. The program memory 17 stores a computer program 18 being adapted for performing the steps of a method according to an embodiment of the present invention when being executed under the control of the processor unit 12. Thereby, in accordance with the various steps of an embodiment of a method of the invention as outlined above for Figs. 3-7, the networked device 10, under the control of processor unit 12 executing computer program 18 is capable of transparently switching network traffic from interface If_B 2 to interface If_A 1. Although the present invention has been described in connection with the preferred embodiments, it is not intended to be limited to the specific form set forth herein. Rather, the scope of the present invention is limited only by the accompanying claims. In the claims, the term comprising does not exclude the presence of other elements or steps. Additionally, although individual features may be included in different claims, these may possibly be advantageously combined, and the inclusion in different claims does not imply that a combination of features is not feasible and/or advantageous. In addition, singular references do not exclude a plurality. Thus, references to "a", "an", "first", "second" etc. do not preclude a plurality. Furthermore, reference signs in the claims shall not be construed as limiting the scope.