Adaptive Routing

Abstract

Methods and devices are shown for improving an existing route between a first node and a second node in a first ad-hoc network that includes a plurality of nodes, the method and devices selecting based on a predetermined performance criterion or criteria one of a plurality of candidate routes between the first node and the second node, wherein each candidate route includes at least one inter-node link that is not included within the first ad-hoc network and at least one candidate route including multiple inter-node links; initializing, via the existing route, the use of the inter-node links of the selected candidate route; and switching from using the existing route to using the candidate route.

Description

FIELD OF THE INVENTION

Embodiments of the present invention relate to adaptive routing. In particular but not exclusively to efficient adaptive routing in low power radio frequency, ad-hoc networks having mobile nodes such as BLUETOOTH® networks.

BACKGROUND TO THE INVENTION

The basic communication network in Bluetooth networks is called a Piconet. The channel access in a Piconet is managed by a Master which is one of the participants in the communication. All other participants in the communication are called Slaves. A master is allowed to serve at most 7 slaves, and a Master can be a Slave in another Piconet. A Slave can also be a Slave in another Piconet. A Master or a Slave that participates in multiple Piconets is called a bridge node and acts as a router between the Piconets. A connected network that spans more than one Piconet is called a Scatternet. Ad hoc multi-hop networks in Bluetooth networks are formed using Scatternets.

Multi hop communications are useful for extending the communication range of a device which is constrained by the maximum device capacity of a Bluetooth Piconet and/or user preferences. For example, a smart phone may prefer to communicate through its trusted laptop or PDA to the outside world. Hence in Bluetooth based ad hoc networks, having two devices within their nominal communication range is not a guarantee that they establish a direct link. This constrains the network topology.

Network topology has a great influence on the performance of the network. A casually designed network topology might be inefficient, and also might preclude some of the applications that otherwise could operate on the network.

Most of the network formation research for Bluetooth based ad hoc networks concentrates on finding an optimal initial topology for a scatternet.

One problem with static network formation protocols is that when a node moves and the topology changes, the initial criteria satisfied by the topology may no longer be valid.

Node mobility calls for a network topology optimization strategy as an integral part of mobility management and/or data traffic routing. This application takes the second approach and addresses the problem of network topology optimization while routing the data traffic.

Adapting the topology of a network based on the requirements of a data flow is known. “A Two-Phase Scatternet Formation Protocol for Bluetooth Wireless Personal Area Network”, Kawamoto, Y. et al, WCNC 2003—IEEE Wireless Communications and Networking Conference, vol. 3, 16-20 Mar. 2003, pages 1453-1458, discloses a two-phase scatternet formation protocol that supports dynamic topology changes. In a first phase, a control scatternet is constructed to support topology changes and route determination. The control scatternet does not operate according to the Bluetooth specification, but instead the piconets are synchronized so that there are common periods when a bridge node switches between piconets without informing the masters about the timing of each switch over. The second phase creates a separate on-demand scatternet whenever a source node wants to initiate data communications with another destination node. The on-demand scatternet formation is itself divided into two steps. In the first step, the control scatternet is used in the selection of a route in the control scatternet between a source and the master serving a destination. In the second step, the master node with which the source node communicates with selects the participating nodes for the on-demand scatternet. The master node uses its own slaves' and adjacent neighbor information to choose the participating nodes. Each master node must maintain all the information of its slaves and bridge nodes within its piconet and adjacent piconets. The on-demand scatternet is torn down when the data transmissions are finished. It will be appreciated that there is a heavy overhead in setting-up and maintaining the control scatternet and that the on-demand scatternet is used only as a temporary addition to the static control scatternet.

BRIEF DESCRIPTION OF THE INVENTION

According to one embodiment of the invention there is provided a method of improving an existing route between a first node and a second node in a first ad-hoc network that comprises a plurality of nodes, the method comprising: selecting based on a predetermined performance criterion or criteria one of a plurality of candidate routes between the first node and the second node, wherein each candidate route comprises at least one inter-node link that is not comprised within the first ad-hoc network and at least one candidate route comprises multiple inter-node links; initializing, via the existing route, the use of the inter-node links of the selected candidate route; and switching from using the existing route to using the candidate route.

According to another embodiment of the invention there is provided a device operable as a first node in a first ad-hoc network comprising a plurality of nodes, including the first node and a second node, and an existing route between the first node and the second node that comprises a plurality of inter-node links, the device comprising: means for selecting based on a predetermined performance criterion or criteria one of a plurality of candidate routes between the first node and the second node, wherein each candidate route comprises at least one inter-node link that is not comprised within the first ad-hoc network and at least one candidate route comprises multiple inter-node links; means for initializing, via the existing route, the use of the inter-node links of the selected candidate route; and means for switching from using the existing route to using the candidate route.

According to another embodiment of the invention there is provided a method of improving an existing route between a first node and a second node in a first ad-hoc network that comprises a plurality of nodes, the method comprising: receiving a message at a device operable as a node of an ad-hoc network; maintaining a cumulative history of the route the message has taken from the first node to the device; maintaining a performance cost of the route the message has taken from the first node to the device; then if a direct link from the device to the second node is available, establishing a candidate route for routing packets between the first and second nodes that comprises the route the message has taken from the first node to the device and the link between the device and the second node, else forwarding the message to at least another device.

According to another embodiment of the invention there is provided a device operable as a node in an ad-hoc network that comprises a plurality of nodes, the device comprising means for receiving a message from a first node; means for maintaining a cumulative history of the route the message has taken to the device; maintaining a performance cost of the route the message has taken to the device; means for determining if a direct link from the device to a second node is available; means for establishing a candidate route for routing packets between the first and second nodes that comprises the route the message has taken from the first node to the device and the link between the device and the second node if a direct link from the device to the second node is available; and means for forwarding the message to at least another device if a direct link from the device to the second node is not available.

According to another embodiment of the invention there is provided a device operable as a node in a first ad-hoc network comprising a plurality of nodes that are interconnected by a plurality of inter-node links, comprising: means for maintaining a list of neighboring devices, wherein each neighboring device is operable as a node in an ad-hoc network and operable to have a direct link to the device and wherein the list comprises a plurality of entries each of which is associated with a neighboring device and each of which comprises an identifier of the associated neighboring device and a parameter indicating the quality of a direct link between the associated neighboring device and the device.

According to another embodiment of the invention there is provided a device, operable as a node in an ad-hoc network that comprises a first link between the device and a first node, comprising: means for identifying when the first link is unused; means for determining whether there is a route via the ad-hoc network other than the first link between the device and the first node; and means for terminating the first link if there is a route.

Some embodiments of the invention provide a protocol that dynamically and permanently changes the network topology to make it efficient for data traffic while satisfying the requirements of the existing data traffic, and honoring the capability constraints of the devices.

A clear advantage of embodiments of the invention over static protocols is better network resource utilization, and mobility management.

BRIEF DESCRIPTION OF THE DRAWINGS

For a better understanding of the present invention reference will now be made by way of example only to the accompanying drawings in which:

FIG. 1 illustrates a process for measuring for each discovered potential neighbor the performance of the respective link to that neighbor and the storage of the results in a probable neighbor list.

FIG. 2 illustrates a start portion of a process. The process determines candidate routes between the source node and a destination node. The start portion occurs at the source node.

FIG. 3 illustrates a process at an intermediate node between the source node and destination node that continues the network optimization process by either establishing a direct link to the destination node or passing the task of establishing a direct link to the destination node to another intermediate node.

FIG. 4 illustrates a process at the destination node that includes the selection of a candidate route and the start of initializing, via the existing route, use of the inter-node links of the selected candidate route.

FIG. 5 illustrates a process at a node within the selected candidate route that continues initializing, via the existing route, use of the inter-node links of the selected candidate route.

FIG. 6 illustrates a process at the source node that starts the switching of the existing route to the candidate route.

FIG. 7 illustrates a process at an intermediate node in the selected candidate route that continues the switching of the existing route to the candidate route.

FIG. 8 illustrates a process at the destination node that terminates the switching of the existing route to the candidate route.

FIG. 9 illustrates a device operable as a node in an ad-hoc network and operable to carry out each of the processes illustrated in FIGS. 1 to 8.

DETAILED DESCRIPTION OF EMBODIMENTS OF THE INVENTION

The proposed network formation protocol operates in two distinct phases. In the first phase an initial network is formed and in the second phase, the network is continuously changed according the traffic needs. The change in the network is a permanent change and there is no automatic reversion to an original or controlling network topology.

The proposed network formation protocol does not describe in detail initial network formation. The initial network can be formed using any of the existing protocols. It might be advantageous to form the network in a shortest possible time in order to make the network operate as quickly as possible and also because the network would be optimized as the data begins to flow in the network.

The adaptive network formation (ANF) is handled by a protocol subsystem residing in each of the nodes participating in the Scatternet. FIG. 9 illustrates a device operable as a node in an ad-hoc network and operable to carry out each of the processes illustrated in FIGS. 1 to 8.

The Adaptive Network Formation process consists of a discovery process as illustrated in FIG. 1 and a network optimization process as illustrated in FIGS. 2 to 8.

The discovery process is responsible for maintaining information regarding potential neighbors, and the quality of the links a node can establish with them.

The network optimization process is responsible for forming new links and tearing down old links based on the data traffic requirements. This process utilizes the information gathered in the discovery process.

The network optimization process comprises:

• • a) the generation of candidate routes between a source node and a destination node as illustrated in FIGS. 2 and 3;
  • b) the selection of a candidate route as illustrated in FIG. 4;
  • c) initializing, via the existing route, the use of the inter-node links of the selected candidate route as illustrated in FIGS. 4 and 5; and
  • d) the switching of the existing route to the candidate route as illustrated in FIGS. 6, 7 and 8.

There is also periodic scanning of the active links, and the pruning of the links that are not recently used while ensuring that the network is not partitioned.

Until the network topology is adjusted by the adaptive network formation method, and the routing tables at the nodes are updated to reflect the modified topology, data is transferred over the current network, on existing routes using the usual data routing functions. This ensures less latency for the data traffic and the compatibility of the network optimization protocol with any routing system.

Referring to FIG. 1, after the initial network is formed, each of the nodes notes its current role: master, slave or bridge node. Each node devotes part of its active time to the Device Discovery process (step 100). This process is described in detail in the Bluetooth Specification. During the process a node switches between inquiry and inquiryscan randomly so as to discover potential neighbor nodes and to be discovered as a potential neighbor node. The amount of time the nodes spend in Device Discovery determines the speed with which the network can adapt to the data needs, and it should be ideally dependent on each node's role, traffic needs, and obligations to the current Piconet or Scatternet structure. Nodes in the Scatternet do not synchronize their discovery processes.

A free node spends all of its time in inquiryscan. Whenever it replies to an inquiry probe it immediately goes into the pagescan (until a connection is established or a timeout occurs) in order that the discovering nodes can connect to it. A connected node need not immediately enter pagescan.

When a node discovers a potential neighbor, it immediately tries to page (step 102) the node. If the discovered node is a free node, then the connection is immediately established (step 104), and the free node joins the Scatternet.

If the discovering node fails to connect to the discovered node (common with nodes that are already connected to the network), it calculates a rendezvous time (step 106) and invites (step 108) the discovered node for a link set-up at the rendezvous time by sending LINK-REQ packet through the existing route in the Scatternet. The packet contains the information as illustrated in Table 1.

LINK-REQ
Protocol id,
Type,
Source,
Destination,
Sequence number,
Approximate rendezvous time,
Duration of rendezvous,
Elapsed time (approximate) from the time request
was originated,
Native clock value of the source
Preferred role of the source in the future link.

Times are measured in slots. The rendezvous time, and rendezvous duration should be approximately calculated based on the expected Scatternet diameter, and should be large enough to accommodate the relay delays, inter Piconet scheduling delays, and the clock drifts of the nodes on the paths.

Each intermediate node that relays the LINK_REQ packet should update the elapsed time field. If elapsed time were greater than half of the rendezvous time, the node would drop the packet, and returns a LINK-REQ-TIMEOUT packet to the source. When the source receives (step 110) the LINK_REQ_TIMEOUT packet it recalculates (step 112) the rendezvous time, and resends (step 108) the packet.

The destination node that receives a link formation invitation (LINK_REQ) should send a reply (LINK_RE_REP) packet. This packet either confirms the acceptance of the rendezvous (possibly with a modified time), or a rejection of the rendezvous. The packet format is the same, and the times are with respect to request origin time. In addition, when a node receives LINK-REQ packet, it updates its probable neighbor list.

A probable neighbor list is maintained by each node. It lists the probable neighbors of that node. It includes information that allows a probable neighbor to be paged i.e. its Bluetooth Clock value and its Bluetooth address and also contains link quality metric(s) for the link to that probable neighbor. The link quality metric(s) may include one or more of the Bit Error Rate (BER) for the link, the bandwidth (BW) of the link etc.

The probable neighbor list maintained by each node may contain the following information.

Probable Neighbor List:
For each probable neighbor there is stored:
BT_ADDR,
BT clock,
Link_quality information,
Time of last updating,
Next updating time
Node role (master or slave) in the temporary link,
Current role of the neighbor

If a node receives (step 114), (step 116) a rejection it again invites the discovered node for a temporary link set up by sending LINK-REQ packet after LINK-TIMEOUT duration.

If the received LINK_RE_REP accepts (step 118) the rendezvous, then at the agreed time (or, if appropriate the newly suggested time) the source establishes (step 120) a temporary link with the discovered node. A temporary link is a link that is established for the sake of measuring (step 122) link quality metrics which are used to update the node's probable neighbor list.

The purpose of temporary link formation is to measure the link quality metrics like Bit error rate (BER), bandwidth (BW), and any other parameters of interest. For example, the information regarding the node's current participation in different Piconets (nodes' role, amount of time spent in each Piconet) can have an affect on the link performance, and hence can also be exchanged during temporary link setup.

When the link is setup, both the nodes also negotiate (step 124) on the future times at which the link is again formed. Therefore, the link quality measurements are done-periodically.

The temporary link is tore down (step 126) as soon as the measurements are done, and the necessary information is exchanged.

If a node is unable to connect to a potential neighbor node in a fixed number of trials for refreshing the link quality metrics, it deems its neighbor has moved, and deletes all the information with respect to this neighbor from its probable neighbor list. When a node loses all of its active links to the neighbors, and is unable to establish a link to at least one of its potential neighbors, it declares itself a free node, and devotes all of its time for inquiry scan, followed by pagescan.

Referring to FIG. 2, when a data connection needs (step 200) to be established with a destination node, a network optimization request is optionally sent to the adaptive network formation module in the node. The network optimization request can be as simple as finding a minimum hop route or a complex requirements on the BER, delay or BW of the route.

If the node already has an active data flow (step 202) to that destination, the request is dropped because supporting multiple flows requires flow or session based routing in the system. We do not assume in this example that the routing function supports this type of routing.

The data source node starts the adaptation process by sending (step 210) or (step 212), (step 214) a BRREQ packet towards the data destination.

If the data destination is a probable neighbor of the data source node (step 204) and if there is available a good enough quality link from the data source to the data destination (step 206) and the link formation is made possible by the current roles of the nodes (step 208), then the data source node would send (step 210) a BRREQ packet via the existing route to the data destination.

If the data destination is not a probable neighbor of the data source node (step 204) or if the data source node does not contain good quality link to the data destination (step 206) or if the link formation is made impossible by the current roles of the nodes (step 208), then the data source node would send (step 212) a BRREQ packet to the next hop node in the current route, and also send (step 214) the BRREQ packet to the probable neighbors with link quality greater than the requested value.

The BRREQ packet contains the following fields:

BRREQ
Protocol id,
type,
Data source,
Data destination,
RouteRequirement,
*PartialRouteMesurement,
*PartialSourceRoute of nodes (in the yet to be formed
network,)
with
their preferred role in a new network,
their infeasible role in a new network, and
the time at which BRREQ was received.
*the time to live in hops.

The fields marked with * are updated at each intermediate node in the route from data source to data destination. The new route is established from data destination to data source using the information accumulated in the BRREQ.

The time stamp information recorded in the BRREQ is used for determining the approximate page or pagescan time while constructing the network in a reverse direction as explained below.

The RouteRequirement value conveys the performance requirement of the new route, and the PartialRouteMeasurement value reflects the cumulative partial route measurement made by the intermediate nodes between the source node and the node that sent the BRREQ. The RouteRequirement and PartialRouteMeasurement values would be used in evaluating the new topology or route.

The infeasible role information is necessary to ensure that the new network topology does not require impossible configurations.

In order to improve the system performance, the preferred roles of the nodes are honored as far as possible.

Referring to FIG. 3, upon receiving (step 300) the BRREQ, an intermediate node checks (step 302) to see if the node already has data traffic from a different source flowing to the data destination. If yes, then it would ignore the BRREQ in order to avoid having multiple routes to the same destination. The intermediate node may also drop the BRREQ, if it cannot form any new links (step 304) for capacity reasons, role restriction reasons, or security reasons. After receiving BRREQ, it notes down (step 306) the BRREQ-id and PartialRouteMeasurement. BRREQ-id is a combination of Data Source, Data Destination, and Sequence Number.

At steps (step 308), (step 310) the intermediate node drops any duplicate BRREQs it receives later except if the duplicate BRREQ packet's PartialRouteMeasurement is better than that of the best previously received BRREQ. The PartialRouteMeasurement of the BRREQ is updated (step 312) with the quality information of the link (step 314) through which the packet is received at the intermediate node. The PartialSourceRoute field of the BRREQ is updated (step 316) with the intermediate node address, and its preferred and infeasible roles. The BREQ packet is stamped (step 318) with the time it was received at the intermediate node.

The intermediate node then checks to see if the data destination is a probable neighbor (step 320) with an acceptable link quality (step 322). If not then the updated BRREQ is forwarded to the nodes that are probable neighbors with good link quality (step 330), and, if the current node lies on the current route to the data destination (step 326), to the next hop node in the current route (step 328). If the data destination is a probable neighbor (step 320) with an acceptable link quality (step 322), then the BRREQ is sent (step 324) via the existing route to the data destination.

Referring to FIG. 4, the data destination node waits (step 400) for some time after it receives the first BRREQ packet in order to collect BRREQ packets from other nodes so as to decide (step 402) on the best new route.

Each received BREQ defines a candidate route in the field PartialSourceRoute and defines a cost for that candidate route in the field PartialRouteMeasurement.

The new route can be selected (step 402) from the candidate routes using, for example, the information in PartialRouteMeasurement and/or PartialSourceRoute. The selection of the route also selects the BRREQ having a PartialRouteMeasurement that defines the new route.

The future neighbor of the data destination node in the new route is determined (step 404) using the PartialSourceRoute associated with the selected BRREQ.

The future page/pagescan timing for the data destination node and the future neighbor node is determined (step 406) from the time stamp information in the selected BRREQ.

The data destination accepts the link formation request BRREQ if its current commitment to the Scatternet, and Piconet are not violated, and sends (step 408) a reply packet BRREP to its future neighbor over the current route. The BRREP is used to establish a new link from the data destination to its future neighbor link. The BRREP is forwarded on towards the data source and enables the formation of links from data destination to data source using the information accumulated in the BRREQ.

The BRREP would contain following fields:

BRREP
Protocol id,
type,
BRREP source,
BRREP destination,
sequence no,
SourceRoute,
Page or Pagescan time,
preferred role.

The SourceRoute is the PartialSourceRoute of the selected BRREQ i.e. the selected candidate route.

At page or pagescan time recorded in the BRREP, both the BRREP source and BRREP destination nodes (neighboring nodes in the new route) enter page or pagescan process to establish the link between them (step 410).

Referring to FIG. 5, the future neighbor receives (step 500) the BRREP and, as the current node, it determines the next hop (future neighbor) towards the data source, and replaces the page/pagescan timing information in the BRREP packet before forwarding it to that node through the existing route. It then establishes a link to the future neighbor, if necessary. This process occurs at every node on the selected candidate route, until the new route is established from the data destination node to the data source node.

The future neighbor of the current node in the new route is determined (step 502) using the SourceRoute of the received BRREP. The future page/pagescan timing is determined (step 504) from the time stamp information in the SourceRoute of the BRREQ. The timestamp information contained in the SourceRoute is used for estimating the approximate time at which the future neighboring node might receive the BRREP, and hence to specify the page/pagescan times. The current node then forwards (step 506) the BRREP packet to its future neighbor over the current route after updating its page/pagescan timing.

If (step 508) a link does not exist between the current node and the adjacent node towards the data destination in the selected candidate route, then the current node determines (step 510) its future pagescan timing and role from the received BRREP and enters (step 512) a page/pagescan mode at the appropriate time to form a link to the adjacent node, towards 5 the destination, in the selected candidate route. The adjacent node is the BRREP source in the received BRREP i.e. the node from which BRREP was forwarded.

Referring to FIG. 6, after receiving (step 600) a BRREP from the data destination, the data source node sends (step 602) a route confirmation BRREF packet to the data destination. A BRREF packet contains the following fields:

BRREF
Protocol id
type
Source
Destination
sequence no
source route
time to live.

Immediately after sending BRREF packet, the data source node updates its routing table and starts to divert (step 604) the traffic to the data destination through the new selected candidate route.

Referring to FIG. 7, when a BRREF packet is received (step 700) by an intermediate node in the new selected candidate route, the node forwards (step 702) the BREF packet to the next node in the selected candidate route and updates (step 704) the routes in its routing table to both data source and data destination nodes from the source route information in the packet.

Referring to FIG. 8, when the data destination receives (step 800) the BRREF from the new route, it updates (step 802) its routing table to the data source node, and starts diverting the reverse traffic to the data source via new route. This completes the adaptation of the network to the needs of the data traffic.

If the data destination is unable to accept the BRREQ, it sends BRREP _NEG to a new neighbor node based on the quality information recorded in the BRREQs it receives. The above topology optimization process starts from this new neighbor instead of the data destination.

The Network Optimization Process at each node periodically checks for unused active links to its neighbors. If a link is not used for data traffic in the last LINK_UNUSED_TIMEOUT period, then it tries to terminate the link, if there is at least one alternative route to the neighbor on that link. In order to determine the existence of an alternate route, it broadcasts a DUPROUTE packet to its neighbors within k hops (determined by the TimeToLive field in the PDU). DUPROUTE packet contains the following fields: Protocol id, type, source, destination, sequence no, Destination_Bluetooth_address, TimeToLive. Each node when it receives the DUPROUT packet, it checks to see if it is a duplicate packet. If it is a duplicate packet, it is not processed further. Otherwise it checks to see if the destination is its immediate neighbor. If not, it decrements the TimeToLive field in the DUPROUTE packet, and broadcasts it to its neighbors. If the destination is its immediate neighbor, then it sends a DUPROUTEREPLY packet to the source. The DUPROUTEREPLY packet contains following fields: Protocol id, type, source, destination, sequence no, current node's id. When a node receives DUPROUTEREPLY packet, it checks for source, destination, and sequence number to identify the accessibility of destination from the source. If it has already seen DUPROUTEREPLY that implies same reachability, it will not further process the packet. Otherwise, if the node is not the originator of the corresponding DUPROUTE query, then it forwards the packet towards the source. If it is the originator, then it terminates the link to the destination, and moves the destination to the probable neighbor list. It also resets the LINK_UNUSED_TIMEOUT value for the link through which it received the original DUPROUTEREPLY packet and thus it can keep that link active for a longer period of time.

If a node loses active links to all of its neighbors, it tries to establish links to its probable neighbors by sending BRREQ with destination address as the probable neighbor, and with no route quality requirements. This message is sent when the probable neighbor nodes try to form a temporary link for quality measurement and for updating the probable neighbor node list. If it fails to establish a link to any of its probable neighbors within a NET_LOST time out period, it declares itself detached from the Scatternet, and tries to get reconnected to the network as a tree node.

When nodes move they can discover the nodes that are within their radio range because every node runs a discovery process. A free node devotes all of its available time to inquiry scan until it replies to inquiry probe. It then enters a pagescan process by which the discovering node can immediately connect it. If the Pagescan procedure does not result in a link within a Pagescan_Timeout period, it terminates its pagescan process, and start behaving like a tree node.

FIG. 9 illustrates a electronic device 10 that is Bluetooth enabled. It is capable of operating as a node in an ad-hoc Bluetooth network. It comprises at least a low power radio frequency transceiver 12, a processor 14 and a memory 16. The memory 16 includes computer program instructions 18 which when loaded into the processor 14 control the operation of the device 10. The instructions 18 include program instructions for carrying out the processes illustrated in FIGS. 1 to 8 depending upon the role of the device: source node, intermediate node, destination node. The device is typically mobile.

Thus, for example, the device of FIG. 9 may be operable as a first node in a first ad-hoc network comprising a plurality of nodes, including the first node and a second node, and an existing route between the first node and the second node that comprises a plurality of inter-node links. In such a case, the device should be understood as including: (1) a selection element or means for selecting based on a predetermined performance criterion or criteria one of a plurality of candidate routes between the first node and the second node, wherein each candidate route comprises at least one inter-node link that is not comprised within the first ad-hoc network and at least one candidate route comprises multiple inter-node links; (2) an initializing element or means for initializing, via the existing route, the use of the inter-node links of the selected candidate route; and (3) a switching element or means for switching from using the existing route to using the candidate route.

Or, the device of FIG. 9 may be operable as a node in an ad-hoc network that comprises a plurality of nodes, the device comprising: (1) a receiving element or means for receiving a message; (2) a route history element or means for maintaining a cumulative history of the route the message has taken to the device; (3) a performance cost element or means for maintaining a performance cost of the route the message has taken to the device; (4) a determiner or means for determining if a direct link from the device to the second node is available; (5) a route establishment element or means for establishing a candidate route for routing packets between the first and second nodes that comprises the route the message has taken from the first node to the device and the link between the device and the second node if a direct link from the device to the second node is available; and (6) a forwarding element or means for forwarding the message to at least another device if a direct link from the device to the second node is not available.

Or, the device of FIG. 9 may be operable as a node in a first ad-hoc network comprising a plurality of nodes that are interconnected by a plurality of inter-node links, the device comprising a neighbor list maintenance element or means for maintaining a list of neighboring devices, wherein each neighboring device is operable as a node in an ad-hoc network and operable to have a direct link to the device and wherein the list comprises a plurality of entries each of which is associated with a neighboring device and each of which comprises an identifier of the associated neighboring device and a parameter indicating the quality of a direct link between the associated neighboring device and the device.

Or, the device of FIG. 9 may be operable as a node in an ad-hoc network that comprises a first link between the device and a first node, the device comprising: an identifying element or means for identifying when the first link is unused; (2) a determining element or means for determining whether there is a route via the ad-hoc network other than the first link between the device and the first node; and a terminating element or means for terminating the first link if there is a route.

Similar variations of the device of FIG. 9 are of course possible and will be evident to any person of skill in the art in view of the teachings hereof.

Although embodiments of the present invention have been described in the preceding paragraphs with reference to various examples, it should be appreciated that modifications to the examples given can be made without departing from the spirit and scope of the invention as claimed. For example, embodiments of the invention may find application in mobile cellular radio telephone networks in which base stations are wirelessly linked to each other.

While endeavoring in the foregoing specification to draw attention to those features of the invention believed to be of particular importance it should be understood that the Applicant claims protection in respect of any patentable feature or combination of features hereinbefore referred to and/or shown in the drawings whether or not particular emphasis has been placed thereon.

In the claims, means-plus-function clauses are intended to cover the structures described herein as performing the recited function and not only structural equivalents, but equivalent structures.

Claims

1. A method of improving an existing route between a first node and a second node in a first ad-hoc network that comprises a plurality of nodes, the method comprising: selecting based on a predetermined performance criterion or criteria one of a plurality of candidate routes between the first node and the second node, wherein each candidate route comprises at least one inter-node link that is not comprised within the first ad-hoc network and at least one candidate route comprises multiple inter-node links,initializing, via the existing route, use of the inter-node links of the selected candidate route, andswitching from using the existing route to using the candidate route.

2. A method as claimed in claim 1, wherein the performance criterion or criteria relate to any one or more of: the number of inter-node links within the candidate route; the bit error rate for the candidate route; the delay for the candidate route; power consumption; interference.

3. A method as claimed in claim 1, further comprising: receiving a plurality of messages routed from the first node to the second node each of which comprises a distinct one of the plurality of candidate routes.

4. A method as claimed in claim 3, wherein the selecting, initializing, switching and receiving occur at the second node.

5. A method as claimed in claim 3, wherein each of the messages received is a BRREQ packet comprising a candidate route defined by a field PartialRouteMeasurement.

6. A method as claimed in claim 3, wherein the selecting uses timestamp information from each of the received messages.

7. A method as claimed in claim 1, wherein the step of selecting uses roles of nodes in the candidate route.

8. A method as claimed in claim 1, wherein the initializing comprises: sending from the second node via the existing route to a neighboring node that neighbors the second node in the selected candidate route, an initialization message for establishing an inter-node link between the second node and the neighboring node.

9. A method as claimed in claim 8, wherein the initialization message comprises timing information for coordinating the neighboring node and second node in establishing an inter-node link.

10. A method as claimed in claim 9, wherein the timing information substantially synchronizes a page scan state and a page state in the second node and neighboring node.

11. A method as claimed in claim 8, wherein the initialization message is a BRREP packet.

12. A method as claimed in claim 1, wherein the step of initializing comprises: sending from the second node an initialization message that is sent within the existing network from node to node of the selected candidate route in sequence and is updated at each node.

13. A method as claimed in claim 12, wherein the initialization message initializes establishment of inter-node links within the selected candidate route where necessary.

14. A method as claimed in claim 12, wherein when the initialization message is received at the first node in the selected candidate route that node switches from using the existing route to using the selected candidate route.

15. A method as claimed in claim 1, wherein the switching from using the existing route to using the selected candidate route, occurs at the second node, when it receives a switch message from the first node indicating that the first node has switched from using the existing route to using the selected candidate route.

16. A method as claimed in claim 15, wherein the switch message is a BRREF packet.

17. A method as claimed in claim 3, wherein the plurality of messages originate when the first node is unable to establish a link of sufficient quality to the second node to transfer data.

18. A device operable as a first node in a first ad-hoc network comprising a plurality of nodes, including the first node and a second node, and an existing route between the first node and the second node that comprises a plurality of inter-node links, the device comprising: a selecting element for selecting based on a predetermined performance criterion or criteria one of a plurality of candidate routes between the first node and the second node, wherein each candidate route comprises at least one inter-node link that is not comprised within the first ad-hoc network and at least one candidate route comprises multiple inter-node links;an initializing element for initializing, via the existing route, use of the inter-node links of the selected candidate route; anda switching element for switching from using the existing route to using the candidate route.

19. A method of improving an existing route between a first node and a second node in a first ad-hoc network that comprises a plurality of nodes, the method comprising: receiving a message at a device operable as a node of an ad-hoc network;maintaining a cumulative history of a route the message has taken from the first node to the device;maintaining a performance cost of the route the message has taken from the first node to the device;then if a direct link from the device to the second node is available, establishing a candidate route for routing packets between the first and second nodes that comprises the route the message has taken from the first node to the device and the link between the device and the second node, else forwarding the message to at least another device.

20. A method as claimed in claim 19, comprising: establishing a candidate route for routing packets between the first and second nodes only if the direct link is available and of sufficient quality.

21. A method as claimed in claim 19, further comprising adding a timestamp to the received message at the device.

22. A method as claimed in claim 19, wherein establishing a candidate route comprises communicating the candidate route to the second node.

23. A method as claimed in claim 22, wherein the candidate route is communicated by forwarding the message to the second node.

24. A method as claimed in claim 19, wherein the device is a node of the first ad-hoc network

25. A method as claimed in claim 19, wherein the at least another device comprises probable neighbors of the device operable as a node of an ad-hoc network.

26. A method as claimed in claim 19, wherein the at least another device comprises at least another device to which the device operable as a node of an ad-hoc network can establish a link of sufficient quality.

27. A method as claimed in claim 19, wherein the at least another device comprises a node adjacent the device operable as a node of an ad-hoc network in the existing route.

28. A device operable as a node in an ad-hoc network that comprises a plurality of nodes, the device comprising: a receiver for receiving a message;a route history maintenance element for maintaining a cumulative history of a route the message has taken to the device;a performance cost maintenance element for maintaining a performance cost of the route the message has taken to the device;a determining element for determining if a direct link from the device to the second node is available;a candidate route establishing element for establishing a candidate route for routing packets between the first and second nodes that comprises the route the message has taken from the first node to the device and the direct link between the device and the second node if said direct link from the device to the second node is available; anda forwarding element for forwarding the message to at least another device if said direct link from the device to the second node is not available.

29. A device operable as a node in a first ad-hoc network comprising a plurality of nodes that are interconnected by a plurality of inter-node links, comprising: a neighbor list maintaining element for maintaining a list of neighboring devices, wherein each neighboring device is operable as a node in an ad-hoc network and operable to have a direct link to the device operable as a node in said first ad-hoc network and wherein the list comprises a plurality of entries each of which is associated with a neighboring device and each of which comprises an identifier of the associated neighboring device and a parameter indicating the quality of a direct link between the associated neighboring device and the device operable as a node in said first ad-hoc network.

30. A device, operable as a node in an ad-hoc network that comprises a first link between the device and a first node, comprising: an identifying element for identifying when the first link is unused;a determining element for determining whether there is a route via the ad-hoc network other than the first link between the device and the first node; anda terminating element for terminating the first link if there is a route.

31. (canceled)

32. A device operable as a first node in a first ad-hoc network comprising a plurality of nodes, including the first node and a second node, and an existing route between the first node and the second node that comprises a plurality of inter-node links, the device comprising: means for selecting based on a predetermined performance criterion or criteria one of a plurality of candidate routes between the first node and the second node, wherein each candidate route comprises at least one inter-node link that is not comprised within the first ad-hoc network and at least one candidate route comprises multiple inter-node links;means for initializing, via the existing route, use of the inter-node links of the selected candidate route; andmeans for switching from using the existing route to using the candidate route.