WO2009046568A1

WO2009046568A1 - Method for forwarding packets via a group of cooperative network elements and network element

Info

Publication number: WO2009046568A1
Application number: PCT/CN2007/002908
Authority: WO
Inventors: Chunyan Yao; Haibo Wen; Jun Zheng; Qingshan Zhang
Original assignee: Alcatel Lucent Shanghai Bell Co Ltd; Alcatel Lucent SAS
Current assignee: Nokia Shanghai Bell Co Ltd; Alcatel Lucent SAS
Priority date: 2007-10-10
Filing date: 2007-10-10
Publication date: 2009-04-16
Anticipated expiration: 2010-04-10
Also published as: CN101822002B; CN101822002A

Abstract

The present invention proposes a method for forwarding packets via a group of adjacent and cooperative network elements in a packet network and a network element for managing the binding information of a mobile node, wherein a mobile node having moved from a first position to a second position and a correspondent node communicate with each other via said group of network elements, said first position and said second position are respectively in the home subnet and a foreign subnet of said mobile node, each of said network elements has both its own address and a specific address of said group which it affiliates to, and comprises a table for storing binding information which is shared among the network elements in said group, said method for forwarding packets comprising: when a network element of said group receives a packet from said correspondent node, the network element will look up the table for the address of the network element currently serving said mobile node; said network element forwards said packet to the network element currently serving said mobile node; and when the network element currently serving said mobile node receives said packet, it looks up the table and forwards said packet directly to said mobile node.

Description

Method for Forwarding Packets via a Group of Cooperative Network Elements and Network Element

Field of the Invention

The present invention relates to the field of communication, and more particularly to forwarding packets in a packet network.

Background of the Invention

The fixed and mobile convergence in edge network is increasingly desirable. In the early deployment of Mobile IP in edge network, locating Home Agent (HA) on Edge Router is a relatively easy and low-cost selection for operators.

Mobile IP is an open standard, defined by the Internet Engineering Task Force (IETF) RFC3344 and RFC3775, that allows users to keep the same IP address, stay connected, and maintain ongoing applications while roaming between IP networks. Mobile IP is part of both IPv4 and IPv6 standards.

Mobile IPv4 is the IETF (Internet Engineering Task Force) standard protocol for handling mobility of an IPv4 node across the Internet. This protocol allows the use of a single fixed IP address regardless of the IP subnet changes, and hence enables continuous reachability for mobile nodes (MN). The fixed IP address is called a home address, and the IP address acquired at each visited network is called a care-of address (COA). The mapping between the home address and the care-of address of a MN is maintained at a special device called a home agent. A home agent intercepts packets on behalf of the MN, encapsulates them and then transmits them to the CoA of the MN by tunneling. A home agent often works together with a foreign agent (FA), which is connected to a foreign network of the mobile network. The HA and the FA are of respectively two types of mobile agent as described in RFC3344, which are defined by IETF.

On the other hand, Mobile IPv6 design and deployment enjoys both the availability of addresses and the extensibility provided by IPv6 protocol. Route optimization signaling enables a mobile IPv6 node to inform its correspondent node (CN) about its new care-of address. This allows both the MN and the CN to send and receive packets using the shortest path between the two. One useful by-product of this feature is location-based services. Mobile IPv6 location update signaling can be used by a CN to infer the geographic location of a MN, and hence provide customized service or content. This optional protocol signaling can be turned off if the MN's location privacy is an issue.

By mechanisms of Mobile IPv4 and Mobile IPv6, a MN can communicate with a Correspondent Node (CN) when the MN is away from the home subnet. When the HA goes down, the MN away from its home subnet cannot communicate properly with the CN. The problem can be illustrated in Figure 1. In Figure 1 , a HA is located in BRAS or edge router, initially the MN is in its home subnet (Position 1 in Figure 1) and is communicating with the CN. When the MN moves to Position 2 (Position 2 in Figure 1), It sends a binding update which provides address map between the home address and the CoA to its HA (Edge Routeri in figure 1). Its HA constructs Binding Information Table (BIT) on it. In Mobile IPv4, the HA of a MN forwards packets from a CN to a MN by a tunnel between itself and the MN or a FA. In Mobile IPv6, when the operation mode between a MN and a CN is in a "bidirectional tunneling" mode, the HA of the MN forwards packets between the CN and the MN by a bidirectional tunnel between itself and the MN. In above two situations, when the HA of the MN (i.e. Edge Routeri) goes down, the communication between the MN and the CN is interrupted. Although multiple HAs can be deployed in a same home subnet, however, the cost would be much higher for operators, and when all HAs in a home subnet fail, the IP packets delivery of all the MNs away from their home subnet but served by the failed HAs is also interrupted.

There is an existing method which can partly solves the problem for IPv6 HA: Deploy multiple HAs in a same home subnet (named Local Recovery Configuration and the recovery process is defined in draft-ietf-mip6-hareliability-02.txt) or different subnets (named Global Recovery Configuration, but no a detailed recovery process is given, only declaimed that the recovery process is similar to Local Recovery Configuration). By this way, N:1 backup can be provided no matter Local Recovery Configuration or Global Recovery Configuration: it means that there is only one active HA working at the same time and the standby HA takes over only when it has detected failure of the active HA. Fig. 2 shows the best existing solution for the problem and its defects. As shown, this existing solution has following disadvantages: firstly, using N:1 backup, the cost is much higher for operators. Secondly, In this solution, the MNs leaving their home subnet are still served by their active HA. When all HAs in the subnet fail (e.g. by hacker/virus attack, outlet circuit/link error or by power outrage in some district), the IP packets delivery of all roaming MNs served by the failed HAs is also interrupted. Another drawback (the third drawback) of this solution is: All HAs in one subnet are one end of the IP tunnels from HA to MN/FA when the MN is roaming (away from its home subnet). Encapsulation /Decapsulation in the HA consume much resource and make the number of roaming MNs served by the HAs limited. The last drawback is that In layer 3 handoff, the roaming MN sends Binding Update message to its HA, and only after the Binding Update message arrives at the HA and has updated the BIT (Binding Information Table), the new path from the HA to the MN can be established, which results in relative long handoff time.

Summary of the Invention

To solve the above problem in the prior art, according to an aspect of the present invention, a method for forwarding packets via a group of adjacent and cooperative network elements in a packet network is proposed, wherein a mobile node having moved from a first position to a second position and a correspondent node communicate with each other via said group of network elements, said first position and said second position are respectively in the home subnet and a foreign subnet of said mobile node, each of said network elements has both its own address and a specific address of said group which it affiliates to, and comprises a table for storing binding information which is shared among the network elements in said group, the method comprises: when a network element of said group receives a packet from said correspondent node, the network element will look up the table for the address of the network element currently serving said mobile node; said network element forwards said packet to the network element currently serving said mobile node; when the network element currently serving said mobile node receives said packet, it looks up the table and forwards said packet directly to said mobile node.

According another aspect of the present invention, a network element for managing the location information of a mobile node is proposed, wherein a mobile node and a correspondent node communicate with each other via a group of adjacent and cooperative network elements, each of said network elements has both its own address and a specific address of said group which it affiliates to, and comprises a table for storing binding information which is shared among the network elements in said group, said network element comprises: a table lookup means for looking up a table for the own address of the network element currently serving the mobile node or looking up the address of said mobile node if said network element is the one which is currently serving said mobile node, upon the reception of a packet from the correspondent node; and a packet forwarding means for forwarding said packet to the network element currently serving said mobile node, or forwarding said packet directly to said mobile node if said network element is the one which is currently serving said mobile node.

Brief Description of the Drawings

These and many other features and advantages of the present invention will become apparent from the following description of the embodiments of the present invention with reference to the drawings, wherein:

- Fig.1 illustrates the problem occurred in a prior art system;

- Fig.2 illustrates the existing solution for the problem and its defects;

- Fig.3 is a flowchart of the method for forwarding packets via a group of cooperative agents according to an embodiment of the present invention;

- Fig.4 shows how an agent group cooperates to forward packets according to the method of the present invention;

- Fig.5 shows how the robustness and load-sharing could be achieved with lower cost than the existing solution, according to the method of the present invention;

- Fig.6 shows how the handoff time could be reduced according to the method of the present invention;

- Fig.7 is a block diagram of the network element according to an embodiment of the present invention. Detailed Description of the Invention

The present invention proposes a method for forwarding packets via a group of adjacent and cooperative network elements in a packet network. This method may be for example applied to the systems shown in Figure 4. With reference to Figure 4, the basic idea of the present invention is illustrated.

As shown in Figure 4, several adjacent (link adjacent or subnet adjacent) HAs can be contracted to form a cooperating group named Agent Group (AG), which group comprises for example at least two adjacent agents. An anycast address is configured in all members of the group, and each member in the group distributes its BIT to others members. Also each member distributes not only the route destined to the network prefix of the subnet served by itself, but also the routes destined to the network prefixes of the subnets served by other members into NSP (Network Service Provider) Network.

Hence, the packet from CN can arrive in any one of the member of the group, ideally the nearest to the CN. The member firstly getting the packet from CN checks its BIT in order to get know to which subnet the MN is now attached and which member serves that subnet. Then it sends the packet from CN to the member currently serving the MN by IP tunneling or by inserting "IP source routing option (for IPv4)7"Type 0 Routing header(for IPv6)". Here tunneling is more preferred. We name the member currently serving the MN as Current Agent (CA). Then CA forwards the packet to the MN.

Each member includes both the IP anycast address configured on it and its IP unicast address acting as HA address in its Agent/Router Advertisement. IP unicast address is included to distribute each other's BIT and to achieve movement detection through subnets served by AG. In the prior art, when the MN moves into other members' subnet, it will send a binding update message to its HA. However, in the method of the present invention, the binding update message is sent to the IP anycast address of the group, and one particular member will receive that message, preferred the CA in the case of the present method. Remember that the destination of a packet with anycast address should be decided by a router, i.e. it is the router who optionally decides to which member in the group that packet is sent. Then, the CA updates its BIT based on the binding update message and informs this update to all other AG members.

When the MN in a foreign subnet sends a packet to the CN, the packet is sent to the CA and is then forwarded to the CN instead of sending it to the MN's HA.

It should be appreciated that, a BIT in each member is created by mobile IP protocol. Each member distributes its BIT to all other members regularly or triggered by request from specific member, or triggered by changes in its BIT. Each of other members receiving the advertisement updates (including adding, deleting and changing) its OBIT. The distribution is delivered by IP unicast with UDP, TCP or RAW IP. In the method according to the present invention, an Over-group Binding Information Table (OBIT) is constructed. The necessary columns in the OBIT including: MN address, COA address and CA address. "CA Address" means to point out which member this binding information comes from.

With reference to Figure 3, the method for forwarding packets via a group of cooperative network elements is described in combination with the system shown in Figure 4.

As shown in Figure 3, firstly, in step 301 , when the MN moves from a first position to a second position, when a network element of the group receives a packet from said CN, said network element will look up its OBIT for the address of the network element currently serving the MN. Herein, as shown in Figure 4, the first position is for example "Position 1", and the second position is for example "Position 2"; M1 is the home agent of the MN, and M2 is the current agent of the MN. As described above, when the MN moves in "Position 2", it will send a binding update message to its CA M2, and then M2 updates its OBIT and informs the update to all other members in the group. For example, when M3 receives a packet from the CN, it will look up its OBIT and know that M2 is the CA of the MN.

Next, in step 302, said network element forwards said packet to the network element currently serving said mobile node. In this embodiment, as M2 being the CA of the MN, M3 will forward the received packet to M2.

Finally, in step 303, when the network element currently serving said mobile node receives said packet, it looks up its table and forwards said packet directly to said mobile node. Herein, M2 looks up it OBIT for the address of MN on receiving said packet, and then forwards said packet directly to the MN.

The above steps relate to the packet from the CN to the MN. It should be noted that, when a CA, for example M2, receives a packet from the MN, it will forward the packet to the CN by means of normal unicast forwarding, which could be implemented by both mechanisms of Mobile IPv4 and Mobile IPv6.

It should be noted that, according to the present method, those members of the group should know the following information by manual or automatic configuration: a. Each member should have an unicast address used to communicate with other members affiliated to the same group. Each member should know the unicast addresses of all other members for intra-group communication. b. Each member should know the ID of the group which it affiliates to, different Group ID identifies different Group. A group ID is in fact a tag used for identifying, for example a number or a sign, which corresponds to the anycast address of the group. c. Each member should know the anycast address of the group which it affiliates to, different anycast address identifies different Group.

As shown in Figure 5, the robustness and load-sharing could be achieved with lower cost than the existing solution, according to the method of the present invention. In Figure 5, the MN is away from its home subnet (e.g., in position 3) and its HA (here M1) fails. In the case that its HA is not in the packet delivery path from the CN to the MN, and for example M3 is the nearest AG member for the CN, Even if M1 fails, the communication between the MN and the CN can continue for M3's serving. In the case that its HA is in the packet delivery path from the CN to the MN, M1's failure results in routing re-convergence in the IP core Network. After the re-convergence completes, the packet delivery path from the CN to the AG will be re-established as CN-M2, and the communication between the MN and the CN can be recovered for M2's and M3's serving. Furthermore, If the MN's home subnet has a backup link connected to M2, not only roaming MNs but also MNs staying at home (where M1 locates) can still communicate with CN if M 1 fails, because M2 can serve them as a new home agent. As conclusion, by the method of the present invention, when the MN's HA is down and the MN is in a foreign subnet, at least the IP packet delivery between the roaming MN and the CN will not be interrupted or can be auto-recovered by the serving of CA instead of HA. This equals that the N+1 backup is provided to the MN's HA. When the MN's HA is running well but the MN is away from its home subnet, the IP packet delivery between the MN and the CN can be served by the CA instead of the HA, this equals that the load-sharing is provided to the MN's HA by other members.

As shown in Figure 6, When the MN moves through the subnets served by the Agent Group, the handoff time could be reduced. In Figure 6, When the MN moves from "Position 2" to "Position 3", the binding update message is sent to the anycast address of the Agent Group and in this case arrives at M3, which triggers M3's OBIT update, and further triggers M3 sending its OBIT update to M2 and M1. When the OBIT update arrives at M2, a new packet delivery path from the CN to the MN can be established as follows: If M1 is the nearest Agent Group member for the CN, the packet sent by the CN arrives at M1 firstly, and then M1 looks up its OBIT using the destination IP address of the packet as key, and finds that the MN is currently served by M2 since the OBIT update from M3 has not arrived at M1 yet. Then the packet is forwarded to M2 by M1 , and M2 checks its OBIT and finds the MN is currently served by M3 since the OBIT update from M3 has arrived at M2. Then the packet is forwarded to M3 by M2, and finally gets to the MN. The newly established packet delivery path is illustrated as the thick line from the CN to the MN in Figure 6. However, in the existing solution, when the MN moves from "Position 2" to "Position 3", a binding update message is sent to its HA (here M1 in Figure 6). Only when the binding update message from the MN gets to its HA and updates the HA's BIT, the new packet delivery path from the CN to the MN can be established. Thus it could be seen that the handoff time of layer 3 is reduced by the method of the present invention.

Thereby, with this method, by providing cooperation among a group of adjacent HAs, both the N+1 backup and the load-sharing for any one of those HAs could be gained with lower cost. Furthermore, when the MN moves through the subnets served by the cooperative HAs, the handoff time could be reduced, and the encapsulation and decapsulation of the IP tunnel packet could be eliminated in the HA. It is therefore an economical and efficient solution for the problem described in the introduction. Based on the same inventive concept, according to another aspect of the present invention, a network element is proposed for managing the location information of a MN. The network element will be described in the following with reference to Figure 7.

Fig.7 is a block diagram of the network element according to an embodiment of the present invention, which is for example is a host or router as the HA, FA or CA. The network element 700 includes a table lookup means 701 and a packet forwarding means 702. When the MN moves from a first position to a second position, for example the first position being "Position 1" and the second position being "Position 2" as shown in Figure 4, and when said network element 700 receives a packet from the CN, the table lookup means 701 looks up the OBIT in the network element 700 for the unicast address of the CA of the MN or looks up the address of the MN if the network element is just the CA of the MN. Then, the packet forwarding means 702 forwards said packet to another network element currently serving the MN, or forwards said packet directly to the MN if said network element is the one which is currently serving the MN. In practice, when the network element 700 receives a packet from the MN, and if this network element is currently serving the MN, the packet forwarding means 702 forwards the packet to the CN by normal unicast forwarding.

In implementation, the network element 700 of this embodiment as well as the table lookup means 701 and the packet forwarding means 702 it includes, may be implemented in software, hardware or a combination of them. For example, those skilled in the art are familiar with a variety of devices which may be used to implement these components, such as micro-processor, micro-controller, ASIC, PLD and/or FPGA etc.. The table lookup means 701 and the packet forwarding means 702 of the present embodiment may be either implemented as integrated into the network element 700, or implemented separately, and they may also be implemented separately physically but interconnected operatively.

In operation, said network element for managing the location information of a mobile node of the embodiment illustrated in connection with Figure 7, may achieve both the N+1 backup and the load-sharing with lower cost than the existing solutions by the cooperation among a group of adjacent network elements. Furthermore, when a MN moves through the subnets served by the cooperative network elements, the handoff time could be reduced, and the encapsulation and decapsulation of the IP tunnel packet could be eliminated in the network element.

Although the exemplary embodiments of the method for forwarding packets via a group of cooperative network elements in a packet network and the network element for managing the location information of a mobile node of the present invention are described above in detail, the above embodiments are not exhaustive, and those skilled in the art can make numerous changes and modifications within the spirit and scope of the present invention. Therefore, the present invention is not limited to those embodiments, the scope of which is defined only by the appended claims.

Claims

1.A method for forwarding packets via a group of adjacent and cooperative network elements in a packet network, wherein a mobile node having moved from a first position to a second position and a correspondent node communicate with each other via said group of network elements, said first position and said second position are respectively in the home subnet and a foreign subnet of said mobile node, each of said network elements has both its own address and a specific address of said group which it affiliates to, and comprises a table for storing binding information which is shared among the network elements in said group, the method comprising:

- when a network element of said group receives a packet from said correspondent node, the network element will look up the table for the address of the network element currently serving said mobile node;

- said network element forwards said packet to the network element currently serving said mobile node;

- when the network element currently serving said mobile node receives said packet, it looks up the table and forwards said packet directly to said mobile node.

2. The method according to claim 1, further comprising:

- the mobile node sends a packet directly to the network element which is currently serving it;

- the network element currently serving said mobile node forwards said packet to said correspondent node by normal unicast forwarding.

3. The method according to claims 1 or 2, wherein said network element looks up the table by using the IP destination address of said packet as keyword.

4. The method according to any of claims 1-3, wherein said table comprising the address of said mobile node, the care of address and the address of the network element which is currently serving said mobile node.

5. The method according to any of claims 1-4, wherein said network element forwards said packet to the network element currently serving said mobile node by IP tunneling or by inserting IP option or routing header.

6. The method according to any of claims 1-5, wherein said network element's own address is an unicast address, and the specific address of said group is an anycast address.

7. The method according to any of claims 1-6, wherein when said mobile node moves from said first position to said second position, it will send a binding update message to said specific address of the group.

8. The method according to any of claims 1-7, wherein said binding update message will be received by the network element currently serving said mobile node, which will update the table based on said binding update message and inform the update to all other network elements in said group.

9. The method according to any of claims 1-8, wherein each of other network elements receiving an advertisement for updating will update the table.

10. The method according to any of claims 1-9, wherein the update includes adding, deleting and changing.

11. The method according to any of claims 1-10, wherein said advertisement for updating is delivered by IP unicast with UDP, TCP or RAW IP.

12. The method according to any of claims 1-11 , wherein said group is composed of at least two adjacent and cooperative network elements.

13. A network element for managing the location information of a mobile node, wherein a mobile node and a correspondent node communicate with each other via a group of adjacent and cooperative network elements, each of said network elements has both its own address and a specific address of said group which it affiliates to, and comprises a table for storing binding information which is shared among the network elements in said group, characterized in that comprising:

- table lookup means for looking up a table for the own address of the network element currently serving the mobile node or looking up the address of said mobile node if said network element is the one which is currently serving said mobile node, upon the reception of a packet from the correspondent node;

- packet forwarding means for forwarding said packet to the network element currently serving said mobile node, or forwarding said packet directly to said mobile node if said network element is the one which is currently serving said mobile node.

14. The network element according to claim 13, wherein if said network element is currently serving said mobile node, said packet forwarding means forwards a packet to said correspondent node by normal unicast forwarding upon the reception of the packet from said mobile node.

15. The network element according to claims 13 or 14, wherein said table lookup means looks up said table by using the IP destination address of said packet as keyword.

16. The network element according to any of claims 13-15, wherein said table comprising the address of said mobile node, the care-of address and the address of the network element which is currently serving said mobile node.

17. The network element according to any of claims 13-16, wherein said packet forwarding means forwards said packet to the network element currently serving said mobile node by IP tunneling or by inserting IP option or routing header.

18. The network element according to any of claims 13-17, wherein after said mobile node have moved from a first position to a second position, it will send a binding update message to said specific address of the group.

19. The network element according to any of claims 13-18, wherein said binding update message will be received by the network element currently serving said mobile node, which will update the table based on said binding update message and inform the update to all other network elements in a group.

20. The method according to any of claims 13-19, wherein each of other network elements receiving an advertisement for updating will update the table.

21. The method according to any of claims 13-20, wherein the update includes adding, deleting and changing.

22. The method according to any of claims 13-21 , wherein said advertisement for updating is delivered by IP unicast with UDP, TCP or RAW IP.

23. The network element according to any of claims 13-22, wherein said first position is in the home subnet of said mobile node, and said second position is in a foreign subnet of said mobile node.

24. The network element according to any of claims 13-23, wherein said network element's own address is an unicast address, and the specific address of said group is an anycast address.

25. The network element according to any of claims 13-24, wherein said group is composed of at least two adjacent and cooperative network elements.