WO2002065731A2

WO2002065731A2 - Macro mobility in communication networks

Info

Publication number: WO2002065731A2
Application number: PCT/NZ2002/000016
Authority: WO
Inventors: Peter Siegfried Meier; Ian Douglas Royds
Original assignee: Tait Electronics Ltd
Current assignee: Tait Electronics Ltd
Priority date: 2001-02-15
Filing date: 2002-02-15
Publication date: 2002-08-22
Anticipated expiration: 2003-08-15
Also published as: NZ509959A; WO2002065731A3

Abstract

Mobile nodes receive services from stationary nodes in a communications network. Primary base nodes have associated groups of peer nodes that are able to transmit all or parts of incoming packets to mobile nodes which remain within geographical zones determined by the groups. A packet addressed to a mobile node is sent from a correspondent node to a care-of node. The packet is forwarded from the care-of node to a group of peer nodes associated with the care-of-node. All or parts of the packet are forwarded from the care-of node and the peer nodes to the mobile node according to location of the mobile node. A group of peer nodes is preferably defined for substantially each node in the network according to expected geographical movement of the mobile nodes.

Description

MACRO MOBILITY IN COMMUNICATION NETWORKS

FIELD OF THE INVENTION

This invention relates generally to packet communication networks and particularly to wireless IP networks in which mobile nodes move in relation to stationary base nodes. Each mobile node generally has a home network and a home address, and can have a variable care-of-address to which packets are forwarded while the mobile node is visiting foreign networks.

BACKGROUND TO THE INVENTION

Packet communication networks such as the Internet are generally based on models involving a hierarchy of layers and protocols, although these are not necessarily named in a consistent way. The Internet, or any IP network, is an inter-network of one or more networks, sometimes referred to as subnetworks. Layer 3 or the network layer, in the case of the Internet, implements a protocol called IP (Internet Protocol) which defines addresses for nodes on a network, and a format for packets, sometimes called datagrams, and routes these packets between nodes on different networks. Layer 4, or the transport layer, is mainly responsible for end-to-end packet ordering, error correction and congestion management, and usually implements connection-oriented or connectionless protocols called TCP (Transmission Control Protocol) and UDP (User Datagram Protocol). Layer 2, or the data link layer, is mainly responsible for transferring data between two adjacent nodes, and optionally involves segmentation of packets from layer 3 into smaller portions, error correction or detection, and management of contention for a shared medium.

Research and recommendations in relation to the Internet have generally been conducted by way of the IETF (Internet Engineering Task Force) which issues documents called Internet Drafts and RFCs (Request For Comments). Several thousand Internet Drafts and RFCs have been published on a wide range of topics. The IETF is also responsible for gradual replacement of IPv4 (IP version 4) that has been used throughout the 1980s and 1990s, by IPv6 which has expanded addressing and routing capabilities, simplification of packet headers, and several other improvements. IP limits the ability of nodes in a network to easily change their points of attachment. Each node with a new attachment must be given a new address. The limitations are inconvenient for modern users who increasingly use mobile equipment such as laptops and packet radios. RFC 2002 from October 1996 describes a mechanism which enables transparent routing of packets to mobile nodes irrespective of their point of attachment, and thereby provide support for mobile IP. An Internet Draft from November 2000 includes specific recommendations for mobile IP in relation to IPv6.

Mobility of nodes between wireless networks has some special problems that have not yet been resolved in a satisfactory way. A mobile node such as a radio can appear to move quickly between the cells of neighbouring base nodes, depending on speed of individual movement and various local effects. As a result, handoffs between base nodes can happen several times during transmission of a single packet and may cause loss or delay of data. Real time transmission of data such as an ongoing voice signal is particularly sensitive to delay. Rapid movements between cells can also cause unduly frequent updates of the care-of-address that must be assigned to the mobile node under IP mobility. Another consequence of rapid mobility is that reception of packets is unduly delayed when they are only forwarded to new locations of a mobile node after the node has moved. It can be an advantage in packet mobility architectures if the next packet to be transmitted to a mobile node is already at a potential new location before the node arrives at that location.

SUMMARY OF THE INVENTION

It is an object of the present invention to provide improvements relating to wireless communication networks in which mobile nodes receive services from stationary nodes, or at least to provide alternatives to existing systems. In general the invention involves primary base nodes having associated groups of peer nodes which are able to transmit all or parts of incoming packets to mobile nodes which remain within geographical zones determined by the groups.

Accordingly in one aspect the invention may broadly be said to consist in a method of assisting movement of mobile nodes in a packet communications network comprising: delivering a packet addressed to a mobile node from a correspondent node to a care-of node, forwarding the packet from the care-of node to a group of peer nodes associated with the care-of-node, and forwarding all or parts of the packet from the care-of node and the peer nodes to the mobile node according to location of the mobile node. Preferably a group of peer nodes is defined for substantially each node in the network according to expected geographical movement of the mobile nodes. The invention also consists in a network which implements methods of this kind. LIST OF FIGURES

Preferred embodiments of the invention will be described with respect to the accompanying drawings, of which:

Figure 1 shows a prior art network system involving a home agent, a foreign agent and a care-of-address in relation to a mobile node,

Figure 2 shows a prior art network system involving a home agent and a co-located care-of-address in relation to a mobile node,

Figure 3 shows a network system having a home agent, a proxy and a care-of-address in relation to a mobile node,

Figure 4 shows a network in which a primary node has a group of associated nodes and proxies in relation to a mobile node, Figure 5 shows transmission cells for a group of nodes which are associated with a primary node according to likely movement of a mobile node, Figure 6 shows movement and handoff of a mobile node between two cells in a group having a radius of one cell,

Figure 7 shows movement and handoff of a mobile node across three cells in a group having a radius of more than one cell, and

Figure 8 shows how packets may be forwarded between primary nodes of different groups of associated nodes.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

Referring to the drawings it will be appreciated that the invention may be implemented in a variety of ways, in a variety of networks, all within the scope of the specification. The systems described here are provided as examples only. Skilled readers will be familiar with related aspects of packet networks and mobile communication so that details from other sources need not be given in full. Some further background information will be given in relation to Figures l and 2.

Figure 1 schematically shows some aspects of networks under IP mobility. Each mobile node 10 such as a laptop computer or other portable communication device, while away from its home network, is associated with a respective home agent 11 in a home network 12. Each home agent is usually a software feature in a local router 13 which connects the network to a wider IP network 14. The wider network can be of any kind and may be connected to still other IP or non-IP networks 15 in various ways. Router 13 is generally responsible for converting addresses from IP to those which may be required on the home network. When operating away from the home network the mobile node 10 generally registers with a foreign agent 16 in a foreign network 17 through a process of advertisement. The foreign agent is again generally a software feature in a router 18 which is local to the foreign network. The foreign agent obtains or provides a care-of-address, updates the home agent regarding the current location of the mobile node, and records the layer 2 address within the foreign network. An IP packet sent to the home address in the home network is forwarded by the home agent to the foreign agent, and from there to the mobile node as represented in the foreign network. Correspondent nodes 19 sending packets to the mobile node may be present anywhere in the system, including the home network, the foreign network, the wider IP network, or other connected networks.

In Figure 1, an association between the home address and a care-of-address is called a Binding for the mobile node. Registration of a Binding optionally takes place by the mobile node or the foreign agent sending an IP packet to the home agent, including Binding Update information which replaces information related to the previous foreign location of the mobile node, if any. The home agent optionally replies with a Binding Acknowledgment. A list of current correspondent nodes is optionally maintained by the mobile node or the foreign agent and these also are optionally informed of the current Binding. Each correspondent node will also optionally provide a Binding Acknowledgment. Packets sent to the mobile node are routed to the mobile node at its care-of-address, either directly by the respective correspondent nodes or indirectly by the home agent, depending on features of route optimisation in particular networks. A packet directed to the home address is generally tunnelled by the home agent by encapsulating the packet in a larger packet directed to the care-of-address. The foreign agent decapsulates the packet and uses layer 2 functions of the foreign network to forward the contents to the mobile node. The routers in each network might act as home or foreign agents for multiple mobile nodes according to the sizes of the networks.

Figure 2 schematically shows other possible aspects of IP mobility. Details of the home network 12 and the wider IP network 14 are generally as before in relation to Figure 1. The mobile node 10 can make a wired or wireless connection to the foreign network 17 as before. In this example however, the mobile node acts as its own foreign agent and maintains a co-located care-of- address within the foreign network. The mobile node is responsible for all mobility functions such as binding updates and decapsulation of packets which are tunnelled from correspondent nodes or a home agent. Co-located care-of- addresses are generally allocated by another component of the wider IP network, typically a dedicated server 20 which operates according to DHCP (Dynamic Host Configuration Protocol). One advantage of co-located addressing is that the separate foreign agents are no longer required and router 18 can be simplified. One disadvantage is an increased requirement for address space in order to provide a separate care-of-address for each mobile node. IPv6 can readily cope with this requirement in comparison with IPv4 and uses only co-located care-of-addresses rather than foreign agents.

Figure 3 indicates one aspect of mobility in accord with the invention. The mobile node 30 is preferably a wireless device which is identified in relation to a home agent 31 in a home network 32. The home agent 31 is generally a software feature in a router 33 in the home network. A wider network 14, other networks 15 and correspondent nodes 19 are generally present as before. Correspondent nodes may be wired or wireless devices, and may be present anywhere in the overall system. The mobile node is able to move to a foreign network 37 having a number of base nodes 35 which might include base stations or base station controllers in a typical mobile radio or telephone network for example. These base nodes can equally reside in the home network of the mobile node. Each base node has a transmission cell within which mobile nodes are served wireless transmissions. Mobile nodes move between cells in accord with handoff procedures as described in more detail below. Each base node is capable of maintaining care-of-addresses for many mobile nodes MN by way of respective proxy systems 36, and might generally be termed a care-of node. Each MN proxy within one base node is either associated with a unique care-of address, or they can all share a care-of address, or some combination of care-of address. Each MN proxy is preferably able to implement specific behaviour related to IP mobility which would normally be associated with a co-located care-of-address or a foreign agent, such as decapsulation and binding updates. A proxy acts on behalf of a respective mobile node within the home or foreign network, and handles substantially all of the IP related signalling by the mobile node to other nodes inside and outside its current network. New proxies are created at base nodes as the mobile node moves around, and old proxies are deleted. A virtual node 38 is effectively formed for each mobile node and appears to other nodes to provide a co-located care-of-address. This allows relatively easy movement of the mobile node in the network.

Figure 4 shows a group 40 of associated base nodes in a foreign network in relation to a mobile node 30. Some or all of these base nodes could optionally reside in the home network of the MN. One node 41 in the group is identified as a primary node that maintains a current care-of-address for the mobile node. Other nodes 42 in the group are associated by virtue only of their proximity to node 41, and do not have to reside in the same subnetwork or administrative domain. The primary node receives packets intended for the mobile node and forwards them to each member of the group by any suitable mechanism, such as IP multicast. The location of the mobile node within the area covered by the group determines which base node currently serves the mobile node. Those base nodes not currently serving the mobile node generally discard the packets that they receive from the primary node. A router 43 is shown by way of example only. There is no particular hierarchy within the group of base nodes, which is flexible in this regard. Together the base nodes in a group provide cells covering a geographical area or zone within which the mobile node will normally remain for a certain period of time, or according to other criteria. Typically the members of a group are determined by the extent of likely or possible movements of a mobile node during a certain period of time, beginning within the cell of the primary node. This period of time, and hence the size of the groups, will be determined by the desired period between updates of the care-of address, and the maximum transmission time of a single packet, whichever is greater. Membership of a group associated with a primary node is usually determined during set up of the network but may also change in time according to actual movements of mobile nodes. Each base node has substantially equal capabilities to form proxies for mobile nodes and is typically a primary node for a group, and an associated node for one or more overlapping groups. The base nodes are generally peers in this regard, although they may differ in ways not directly related to mobility services. Each node in a group is identified by a multicast or group address for reception of packets that are to be transmitted to mobile nodes.

Figure 5 shows an arrangement of cells for a primary node BN1 and a group of associated nodes BN2-BN8, by way of example. The cells are idealised as overlapping circles 50 about respective base nodes but in practice have irregular outlines determined by the local propagation environment. The boundary of each cell represents an area within which a mobile node generally finds acceptable reception for transmissions by the respective base node and can generally receive services from that node. In regions of overlap between cells a mobile node may find equally good reception from several base nodes. Handoff between primary nodes of different groups and between base nodes within a group can occur according to various criteria, such as link goodness measurements, generally determined in layer 2. The geographical extent of each group is intended to reduce the number of binding updates that might be required in relation to primary nodes as mobile nodes move around under typical circumstances. A primary node generally remains the care-of node for a mobile node while the mobile node remains within the coverage area of the group, or zone of the primary node. There is generally no need for a binding update until the mobile node approaches the edge of the zone and this reduces network traffic. A packet sent by the home agent of the mobile node or by a correspondent node is multicast by the primary node to each member of the group, using the group address. As the mobile node moves between cells in the group, all or parts of the packet may be transmitted by different base nodes, and receive separate acknowledgments from the mobile node. Similarly packets transmitted by the mobile node may be received all or in part by different base nodes in a group. Parts received by different nodes are typically assembled at the primary node, generally in layer 2, and forwarded as a complete packet to the intended IP destination. If they are assembled in some other node, the complete packet can optionally be forwarded via the primary node, according to the requirements of IP mobility and IP routing.

In Figure 5, there are two main constraints on mechanisms within a network that decide when to change the primary node and update the care-of-address for a mobile node. Firstly, that Binding Updates to the respective home agent and optionally to any current correspondent nodes should not be sent more often than necessary. Frequent updates add a processing burden in the primary node and the home agent, and add to network traffic overall. The Ipv6 mobility draft currently suggests a maximum initial rate for repeated binding updates with the same care-of address to an individual node, and a reduced ongoing rate. No explicit constraint is provided in relation to updates with new care-of-addresses. Typical rates in relation to updates for new care-of- addresses should be approximately similar to those suggested in the Ipv6 mobility draft. Secondly, the care-of-address should be updated to a new primary node before a mobile node moves outside the zone of the current primary node. Incoming or downlink packets which are received by a primary node that is no longer current or able to serve the mobile node within its associated group may be lost, according to the chosen implementation for removal of out-of-date proxies. The group to which downlink packets are multicast for transmission to a mobile node should track and remain approximately centered on the mobile node. It should not usually be possible for the mobile node to move abruptly out of the zone of cells provided by the group.

Figures 6 and 7 indicate the timing of possible updates of a care-of-address for a mobile node moving on paths 60 and 70 under different circumstances. In Figure 6, the zone has a radius of one cell outside that of the primary node BN1, with only part of the zone having been shown. BN1 is a primary node with a group of associated nodes including BN2, is an associated node of the group having BN2 as the primary node, and is generally a member of several other groups. When the mobile reaches PI the care-of-address is updated with BN1 as the primary node. A handoff to BN2 within the group of BN1 is decided at P2. At P3 the care-of-address is updated to BN2 as the next primary node. At P4 the mobile would lose service if the update from BN1 to BN2 had not taken place. In Figure 7, the group zone has a radius of more than one cell outside the cell of the primary node. BN1 is a primary node having a group which includes BN2 and BN3. At PI the mobile node has a care-of-address provided by BN1. At P2 a handoff occurs to BN2 within the group. At P3 a handoff occurs to BN3 which then becomes the primary node and provides a care-of- address. BN3 can optionally update BN1 regarding the change and usually carries out other mobility functions such as updates to the home agent and any current correspondent nodes. In normal circumstances represented by these figures, the new primary node is part of the group of nodes associated with the previous primary node, it is likely that some packets will be received at the previous primary node with the previous care-of-address, after the mobile node has moved to the zone of a new primary node with a new care-of-address. These packets are forwarded from the previous primary node to its peer group of nodes as normal, one of which nodes will be the new primary node.

Figure 8 shows how packets coming from a home agent or correspondent node may be forwarded between two primary nodes having previous and current bindings in relation to a visiting mobile node. There are two main reasons why a mobile node might not have received a packet while within the group of a previous primary node, generally due to various time delays in processing or transmission of packets. Firstly, the packet might have been sent from the correspondent node to the previous primary node before receipt of the most recent binding update from the new primary node. The correspondent node would have directed the packet to the previous care-of-address rather than the current care-of-address. Secondly, the previous primary node might have held a queue of packets awaiting transmission to members of its associated group, or directly to the mobile node. New bindings can arise in relation to packets which are held in a queue.

In Figure 8, a mobile node MN moves from the zone of a primary node BN1 to the zone of a primary node BN3 and its care-of-address is changed. Each of these stationary nodes belongs to the group of nodes that is associated with the other, and both groups also include another stationary node BN2. A correspondent node CN sends packets PI, P2, P3 to the mobile node, using care-of-addresses according to the most recent binding updates that have been received. There may be many other correspondent nodes transmitting to and receiving from the mobile node at any particular time, but only one has been shown for clarity. A binding update BU is sent to the correspondent node by BN3 at an appropriate time. Many other binding update messages will also normally be sent, including messages to the home agent and to other correspondent nodes of MN, all generally omitted for clarity. Packets are generally received and queued at each node using layer 3 processes, then passed for transmission one by one in layer 2 processes. Packets are generally labelled according to their passage through a particular primary node in the foreign network. The packets may also be segmented in layer 2 into frames or segments having lengths that are acceptable within the foreign network, for example. It will also be appreciated that transmissions to and from the mobile node in relation to the base nodes preferably but not necessarily involve proxy systems as described above.

In the example of Figure 8, packet PI is first sent by CN to BNl which has the current care-of-address for MN. Packet PI is then multicast by BNl as .GPl to BN2 and BN3 which are both members of the group of nodes associated with BNl. Meanwhile MN moves from the zone related to BNl into the zone related to BN3 and acquires a new binding to BN3. BN3 sends the binding update BU to CN, including a new care-of-address for the mobile node. Packet P2 is sent by CN to BNl before receipt of BU. Packet P3 is sent by CN after receipt of BU and is directed to BN3. Packet P2 is multicast by BNl as GP2 to BN2 and BN3. Packet P3 is multicast by BN3 as GP3 to BNl and BN2. These multicast transmissions might involve some disordering and duplication of packets sent and received at particular nodes. Disordered or duplicate packets are readily recognised by their origin within the particular group and reordered within a queue or discarded as appropriate. BN2 readily reorders the first instance of packet P2 received from BNl in GP2, ahead of packet P3 received earher from BN3 in GP3, and discards the second instance of packet P2 received from BN3.

It will be noted that all or parts of a packet can be transmitted or received concurrently or sequentially at one or more base nodes and split or recombined using a variety of micro-mobility, macro-diversity or soft handover mechanisms.

Claims

CLAIMS:

1. A method of assisting movement of mobile nodes in a packet communications network comprising : delivering a packet addressed to a mobile node from a correspondent node to a care-of node, forwarding the packet from the care-of node to a group of peer nodes associated with the care-of-node, and forwarding all or parts of the packet from the care-of node and the peer nodes to the mobile node according to location of the mobile node.

2. A method according to claim 1 further comprising: defining a group of peer nodes for substantially each node in the network according to expected geographical movement of the mobile nodes.

3. A method according to claim 2 further comprising: defining the groups to cover geographical areas within which a mobile node is likely to remain during transmission of a packet to or reception of a packet from the mobile node.

4. A method according to claim 2 further comprising: defining the groups to cover geographical areas such that the rate at which the home agent and/or correspondent nodes are updated is kept below some desired maximum.

5. A method according to claim 2 wherein: each node in the network is a member of one or more groups of peer nodes associated with respective care-of-nodes and may or may not also be a care-of-node with a group of peer nodes.

6. A method according to claim 1 further comprising: transmitting part of a packet to the mobile node from the care-of node and another part of the packet to the mobile node from a peer node of the care-of node.

7. A method according to claim 1 further comprising: transmitting part of a packet to the mobile node from a peer node of the care-of-node, and another part of the packet to the mobile node from another peer node of the care-of-node or from the care-of-node.

8. A method according to claim 1 further comprising: determining the current location of the mobile node, selecting a care-of node with an associated group of peer nodes, and advising a home agent and/or correspondent nodes for the mobile node regarding the care-of node.

9. A method according to claim 8 further comprising: selecting a new care-of node and updating the home agent and/or correspondent nodes only after a predetermined or calculated interval since a previous update.

10. A method according to claim 1 further comprising: selecting a new care-of node according to a change in location of the mobile node and optionally advising a previous care-of-node regarding the new node.

11. A method according to claim 1 further comprising: forwarding all or parts of the packet from the care-of node and the peer nodes to the mobile node by transmission over a wireless medium.

12. A method according to claim 1 wherein: each peer node has substantially equal capability in the network and forms part of one or more groups associated with other nearby nodes.

13. A method according to claim 1 wherein: the care-of-node acts temporarily as a proxy for the mobile node in relation to the network.

14. A method according to claim 1 further comprising: receiving part of a packet from the mobile node at one stationary node and another part at another stationary node, reassembling the complete packet, and forwarding it to a correspondent node from the care-of node.

15. A communications network which implements a method according to any one of the preceding claims.

16. A node for a communications network that implements a method according to any one of the preceding claims.