Wireless communications apparatus, and routing control and packet transmission technique in wireless network

The present invention generally relates to an ad-hoc network consisting of multiple wireless nodes mutually connected via wireless links to structure a self-configured backbone network. More particularly, the present invention relate to routing control for determining the optimum route in the ad-hoc network by taking into account both the node cost and the link cost, and to packet transmission capable of reducing delay at nodes due to relay traffic.

An ad-hoc network is a technique for self-configuring a communications network with multiple communications equipment sets that operates without a specific control station.

A wireless ad-hoc network is adapted to realize the ad-hoc network technique using radio links or wireless links in the mobile environment, and many studies and much research have been conducted regarding to the wireless ad-hoc network.

It is proposed to introduce the concept of a spanning tree, which is generally used in a wired network, into a wireless ad-hoc network when connecting communications apparatuses or access points via wireless links. (See, for example, JP 2000-69046A publication.) It is also proposed to prevent a loop to avoid wasteful packet relay when creating a spanning tree in a wireless network. (See, for example, JP 2000-78147A publication and JP 2003-188811A publication.)

Since ad-hoc networks are established in a self-configuring manner by communications apparatuses, if workload is concentrated on a specific communications apparatus, the influence spreads to all the apparatuses that transmit and receive data via that specific communications apparatus. In the above-described publications JP 2000-69046A and JP 2000-78147A, workload concentrates on a specific wireless base stationbase station located at the root of the tree, and therefore, the throughput of the entire network is reduced.

On the other hand, JP 2003-188811A proposes to maintain the communication quality by not establishing a channel with respect to a wireless base stationbase station with a low receive-power level when creating a tree. However, this method is not capable of calculating a link cost reflecting adaptive changes in communication rate in the wireless environment. Consequently, the throughput of the entire network cannot be improved.

To avoid the concentration of workload on specific equipment, it is proposed to use the workload on a communications apparatus as an indicator when performing routing control in an ad-hoc network. See, for example, Sung-Ju Lee and Mario Gerla, “Dynamic Load-Aware Routing in Ad hoc Networks”, Proc. ICC 2001, June, 2001. Also, see Hakoda, et al. “Performance Evaluation of Mobile Ad Hoc Routing Protocols Based on Link Expiration time and Load of Node”, IEICE Journal, Vol. 85-B, No. 12, 2002.

It is also proposed to switch between a link cost, which represents communication available time between terminals, and a node cost, which represents the busy rate of a terminal, depending on the network traffic. See Takahashi, et al., “Metric Selection Mobile Ad Hoc Routing Protocol Based on Link and Node Costs”, Technical Report of IEICE, NS 2003-80, RCS2003-103 (2003-7) at 67-72.

With the above-described conventional techniques, load distribution cannot be realized in the network. Or, even if workload is dispersed, efficient use of radio resources cannot be realized.

Therefore, it is an object of the present invention to solve the problems in the conventional techniques and to provide a routing control and packet transmission technique for a wireless network that can achieve both workload distribution among nodes and efficient use of radio resources.

The technology aimed by the present invention may fall in between mobile ad-hoc consisting of only mobile terminals without a specific control station and an infrastructure network requiring a control station; however, it is also applicable to mobile ad-hoc. Wireless communications apparatuses form a wireless backbone network in a self-configuring manner, each apparatus behaving like a wireless base station or an access point to control user terminals existing under the communications apparatus.

In order to realize both efficient use of the radio resources and distribution of workload in the network, a node cost is calculated, in addition to a link cost, at each of the wireless communications apparatuses in the network to take the node traffic into account. The node cost can be calculated based on the amount of data accumulated in the buffer or queue time. The calculated node cost is added to the link cost representing the state of the radio environment to obtain the total cost required to reach the destination of a packet. In other words, the packet transmission route is determined taking into account both the node cost and the link cost.

Independent of or in combination with the calculation of the node cost, a wireless communications apparatus (AP) is configured to have a backbone-system buffer for temporarily storing packets received from other nodes (or access points) and an access-system buffer for temporarily storing packets received from user terminals (STA) located under this wireless communications apparatus (AP) separately. Data transmission from these two buffers is scheduled based on a prescribed order of priority so as to reduce delay occurring at a node (wireless communications apparatus or access point) on which traffic is likely to converge.

In one aspect of the invention, a wireless communications apparatus used in a wireless communications network consisting of multiple wireless communications apparatuses mutually connected via wireless links is provided. The wireless communications apparatus comprises:

(a) a buffer configured to temporarily store information received from an adjacent wireless communications apparatus or a user terminal located under the wireless communications apparatus; (b) a routing control information processing unit configured to estimate a node cost representing traffic at the wireless communications apparatus and a link cost representing a radio condition of the link to determine a transmission route based on cost information reflecting both the node cost and the link cost; and (c) a routing control unit configured to designate a next hop to which a data item accumulated in the buffer is to be transmitted according to the determined transmission route.

By determining the transmission route based on both the link cost representing the wireless condition between adjacent wireless communications apparatuses and the node cost representing the traffic at the node, concentration of workload on a specific node can be avoided, while efficiently using the radio resources. Consequently, wireless transmission can be performed more efficiently.

Preferably, the routing control information processing unit is configured to exchange the cost information with the adjacent wireless communications apparatus, and upon receiving arbitrary cost information, to produce new cost information by adding the node cost of this wireless communications apparatus and the link cost from the adjacent wireless communications apparatus and transmit the new cost information to the next hop.

By allowing the wireless communications apparatus to exchange the cost information including the node cost and the link cost with other wireless communications apparatuses in the wireless network, a route can be determined in a self-configured manner, taking into account both the radio environment and the node traffic.

In another aspect of the invention, a wireless communications apparatus, which is used in a wireless communications network consisting of multiple wireless communications apparatuses mutually connected via wireless links, comprises:

(a) a first buffer configured to temporarily store a first information item received from an adjacent wireless communications apparatus; (b) a second buffer configured to temporarily store a second information item received from a user terminal located under the wireless communications apparatus; (c) a priority control unit configured to schedule transmission of the first information item and the second information item according to transmission priority; and (d) a routing control unit configured to determine a transmission route for the first information item and/or the second information item if they are required to be transmitted to a next hop, based on cost information including a node cost representing traffic at the wireless communications apparatus and a link cost representing the radio condition of the link.

By storing information items in two types of buffers depending on where the information items are received from, and by controlling the transmission order of the information items, the packets can be relayed to the next hop without causing a significant delay even at a node on which traffic is apt to converge.

In addition, for the information items needing to be transmitted to the next node, the transmission route is determined based on the cost information taking both the node cost and the link cost into account.

With this arrangement, concentration of workload on a specific node can be avoided, and the overall network efficiency can be maintained high by setting a suitable transmission route avoiding a heavy-traffic node. Consequently, radio resources can be used efficiently.

Other objects, features, and advantages of the invention will become more apparent from the following detailed description when read in conjunction with the accompanying drawings, in which

FIG. 1 illustrates an example of an ad-hoc network structured by multiple access points (wireless communications apparatuses), to which the present invention is applied;

FIG. 2 is a block diagram of a wireless communications apparatus (AP) according to an embodiment of the invention;

FIG. 3 illustrates an example of priority control using segmen buffers;

FIG. 4 illustrates an example of network topology divided into two blocks;

FIG. 5 is a block diagram of a wireless communications apparatus (AP) according to another embodiment of the invention; and

FIG. 6 is a diagram used to explain cost estimation.

The present invention is described in detail below in conjunction with the attached drawings.

FIG. 1 is a schematic diagram illustrating an example of the network configuration of an ad-hoc network including multiple access points (wireless communications apparatus) AP 1 -AP 8 that function as wireless base stations structuring a wireless backbone network. In addition, all or some of the access points AP 1 -AP 8 function as wireless base stations controlling user terminals (STA) located under the associated access points. The access points AP 1 -AP 8 are mutually connected via wireless links to form the wireless backbone network. A user terminal STA 1 is located under the access point AP 4 , and another user terminal STA 2 is connected to the access point AP 5 via a cable in this example. The access point AP 5 and the user terminal STA 2 are integrated with each other to function as a single apparatus AP 5 ′. Although not shown in FIG. 1 , there may be some user terminals existing under each or some of the access points.

In the specification, a “node” or an “access point” means a communications apparatus that has a packet relaying function and behaves as a wireless base station, and a “user terminal” means communications equipment that does not have a packet relaying function. The apparatus AP 5 ′ in FIG. 1 functions as both an access point and a user terminal, and it is, for example, a fixed television set or a mobile terminal with ability of access point. The apparatus AP 5 ′ is a part of the wireless backbone network.

In FIG. 1 , the apparatus AP 1 functions not only as an access point, but also as a gateway to the wired Internet. Adjacent access points connected by the bi-directional arrows transmit and receive data items according to the tree extending from access point AP 1 functioning as the root. The root AP is not limited to access point AP 1 , but two or more transmission trees may be produced dynamically in the network, designating two or more root APs (access points).

When access points AP 6 , AP 7 , and AP 8 transmit or receive data items to or from the wired network, such data items have to pass through AP 4 connected to the gateway point AP 1 . If, in this state, user terminal STA 1 located under AP 4 is transmitting or receiving a large amount of data, congestion occurs at AP 4 , and the data items routed through AP 4 will be delayed greatly.

To prevent such a delay in the wireless network, each of the access points AP 1 -AP 8 is configured to schedule transmission of data items received from other access points (AP) or the user terminal (STA) located under the access point, according to a prescribed order of priority. In addition, each of the access points AP 1 -AP 8 is configured to determine a transmission route, based on the cost value reflecting the node cost representing traffic at that access point and the link cost representing the radio conditions of the wireless link, as necessary.

FIG. 2 is a schematic block diagram of an access point (AP) or wireless communications apparatus 10 according to an embodiment of the invention. The access point 10 has a backbone-system receiving buffer 11 for temporarily storing data items received from other access points 10 A, and an access-system receiving buffer 12 for temporarily storing a data item received from the user terminal (STA) 20 located under the access point 10 . A priority control unit 13 schedules transmission of the data items from these two buffers 11 and 12 , according to prescribed transmission priority. A routing control unit 14 designates transmission routes for the data items accumulated in the receiving buffers 11 and 12 . The routing control unit 14 may be configured so as to carry out cost estimation. In this case, the routing control unit 14 determines the transmission route, as necessary, based on the estimated cost value reflecting both the node cost and the link cost. The details of estimation of the cost value are described below.

The access point 10 also has a backbone-system transmitting/receiving unit 15 for transmitting and receiving data items to and from other access points 10 A, and an access-system transmitting/receiving unit 16 for transmitting and receiving data items to and from user terminals (STA) located under the access point 10 .

The backbone-system transmitting/receiving unit 15 has a determination unit 18 for determining the destination of the received packet (data item). If the packet is to be transmitted to the next node of the backbone network, the data item is stored in the backbone-system receiving buffer 11 . If the received packet is addressed to the user terminal (STA) 20 located under the access point 10 , then the data item is transmitted via the access-system transmitting/receiving unit 16 to the user terminal (STA) 20 , without being supplied to the backbone-system receiving buffer 11 .

Although, in the example shown in FIG. 2 , two separate radio interfaces (backbone-system transmitting/receiving unit 15 and access-system transmitting/receiving unit 16 ) are used, a single interface may be used to implement the embodiment. In this case, the received packets can be sorted to the backbone-system receiving buffer 11 and the access-system receiving buffer 12 by checking the source addresses of the received packets.

Referring to FIG. 1 , the data item transmitted from AP 6 to the Internet is received at the backbone-system transmitting/receiving unit 15 of AP 4 , which node is to relay the data item to the next node, and stored in the backbone-system receiving buffer 11 . There is a user terminal STA 1 that is currently transmitting a large amount of data, which data are stored in the access-system receiving buffer 12 of AP 4 , via the access-system transmitting/receiving unit 16 . Because the backbone-system receiving buffer 11 and the access-system receiving buffer 12 are provided separately, the backbone-system buffer 12 can accept the relay traffic from other access points 10 A without delay, even if the data items to be accumulated in the access-system receiving buffer 12 are about to exceed the capacity of the access-system receiving buffer 12 .

The transmission priority may be given to, for example, data transmission from the backbone-system receiving buffer 11 , or alternatively, the transmission ratio of the backbone-system receiving buffer 11 to the access-system receiving buffer 12 may be set to 2/1 (2 to 1). The transmission priority may be determined in advance by a network manager, or automatically determined in a variable manner based on the past record of data.

By outputting data items received from the backbone network at a prescribed rate, the data items being relayed to the destination are appropriately transmitted to the next node even if a large amount of data is input from the user terminal (STA) 20 located under the access point 10 . With this arrangement, delay occurring at an access point at which traffic is likely to converge can be reduced, while efficiently using the network resources.

At least one of the backbone-system receiving buffer 11 and the access-system receiving buffer 12 may be segment according to the types of the data items.

FIG. 3 illustrates an example of transmission priority for the segment buffer structure. The backbone-system receiving buffer 11 is divided into three portions BB 1 , BB 2 , and BB 3 . The access-system receiving buffer 12 is divided into three portions AB 1 , AB 2 , and AB 3 . The buffer segments BB 1 and AB 1 are used to temporarily store data items, such as audio data, for which transmission delay is not acceptable. The buffer segments BB 2 and AB 2 are used to store data items with less strict delay criteria. The buffer segments BB 3 and AB 3 are used to store data items for which a certain degree of transmission delay is acceptable.

When transmitting data items stored in the segment buffers, transmission priority control is performed. For example, the data items are taken out of the buffer segments BB 1 and BB 2 of the backbone-system receiving buffer and transmitted in this order. Then, the data item stored in the buffer segments AB 1 and AB 2 of the access-system receiving buffer 12 are extracted and transmitted in this order. Then, the data items in the buffer segments BB 3 and AB 3 are taken out and transmitted in this order. Alternately, although not shown in FIG. 3 , the data items may be extracted and transmitted in the order of BB 1 , AB 1 , BB 2 , BB 3 , AB 2 , and AB 3 .

By setting transmission priority using segment buffers, the data items with higher priority can be transmitted in the appropriate order even if such data items are accumulated in the access-system receiving buffer 12 . In other words, data items stored in the access-system receiving buffer 12 may be placed in the transmission buffer ahead of data items in the backbone-system receiving buffer 11 as long as higher priority is given to such data items. This arrangement can prevent data items, for which much delay is not acceptable, from being lost, and consequently, communication quality can be improved.

Returning to FIG. 2 , the access point 10 has a monitoring unit 17 ′ for monitoring the amount of data accumulated in the backbone-system receiving buffer 11 . If the data accumulated in the backbone-system receiving buffer 11 has exceeded or is exceeding the capacity of the buffer, large delay will occur at that node. To prevent this problem, if the data accumulated in the backbone-system receiving buffer 11 have exceeded a prescribed threshold, the monitoring unit 17 transmits a message reporting the occurrence of congestion to the adjacent access points. In response to the congestion message, the adjacent access points rearrange transmission routes so as to get around the congested node. Similarly, when the access point 10 receives a congestion messages from an adjacent access point 10 A, the routing control unit 14 designates the next hop according to the newly arranged transmission routes so as to avoid this node (which has transmitted the congestion message). The rearrangement of the transmission routes at the occurrence of congestion is be described below.

In the above example, workload concentrates at access point AP 4 to which a large a amount of data is currently transmitted from the user terminal STA 1 located under the access point AP 4 . The same applies to access point AP 1 functioning as the gateway to the Internet because all the traffic accessing the Internet has to pass through AP 1 . Similarly, traffic congestion also occurs at the access point located at the boundary between adjacent blocks to link up these blocks in a network divided into multiple blocks consisting of access points.

FIG. 4 illustrates an example of the network configuration divided into two blocks, each block being structured by multiple access points. Although not shown in FIG. 4 , a user terminal (STA) may exist under the access points (AP) in each block. The access points for linking up these two blocks (that is, AP 4 and AP 7 in FIG. 4 ) are always used when transmitting data between user terminals located in different blocks. Accordingly, even if there is little traffic from the user terminal located under the link-up access point (AP 4 or AP 7 ), the backbone-system receiving buffer of the link-up access point is likely to overflow due to data transmission from other access points.

To overcome this problem, a transmission route is rearranged so as to avoid the link-up access point when a data item is addressed to a user terminal (STA) located in the same block.

For example, when transmitting data from AP 9 to AP 10 , there are two different routes, AP 9 →AP 7 →AP 10 and AP 9 →AP 8 →AP 10 . If traffic is converging at the link-up access point AP 7 , it is desired to select the latter route for reducing the workload on AP 7 .

In the radio environment, a bypass route is likely to reduce the transmission rate and degrade the performance of the overall system. Accordingly, it is required to select an alternative route so as to avoid the congested node with as little bypass (detour) as possible. To realize this, the alternative route is selected taking both the node traffic and the radio environment between adjacent nodes (AP) into account in the embodiment.

If a spanning tree protocol standardized in IEEE 802.1D is employed in routing control, each access point writes a cost required from the root node (AP) of the tree in the routing control packet (e.g., bridge protocol data unit: BPDU) transmitted at prescribed intervals.

In this embodiment, a cost value reflecting both a node cost representing the traffic of the node and a link cost representing the radio conditions is written in the routing control packet (BPDU), which is transmitted along the tree in the network. The cost value written in the BPDU may be estimated based on the expression,

Cost = ( cost ⁢ ⁢ representing ⁢ ⁢ the ⁢ ⁢ node ⁢ ⁢ traffic ) × ( cost ⁢ ⁢ representing ⁢ ⁢ the ⁢ ⁢ radio ⁢ ⁢ conditions ) or ⁢

⁢ Cost = ( cost ⁢ ⁢ representing ⁢ ⁢ the ⁢ ⁢ node ⁢ ⁢ traffic ) + ( cost ⁢ ⁢ representing ⁢ ⁢ to ⁢ ⁢ the ⁢ ⁢ radio ⁢ ⁢ conditions ) .

The cost may be normalized by multiplying by an appropriate coefficient.

The node cost representing the node traffic may be determined from the amount of data accumulated in the receiving buffer or the queue time. The link cost representing the radio conditions may be determined from the data transmission rate (link rate) or the condition (including the intensity) of radio waves.

By selecting the node (AP) so as to minimize the cost value, an alternative route with as little bypass as possible can be selected in a self-directed manner, while avoiding a node on which workload is converging.

FIG. 5 is a block diagram illustrating a wireless communications apparatus (or an access point) 30 with a cost estimation function. The access point (AP) 30 has a receiving buffer 31 , a routing control information processing unit 38 , and a routing control unit 34 . The receiving buffer 31 temporarily stores data items received from the user terminal 20 located under the access point 30 and/or another access point 30 A. The routing control information processing unit 38 estimates a node cost representing the traffic at this node and a link cost representing the radio conditions, and determines a transmission route based on the cost value reflecting both the node cost and the link cost. The routing control unit 34 designates the next hop to which the data item stored in the receiving buffer is to be transmitted according to the determined route.

The routing control information processing unit 38 transmits a routing control packet to the adjacent access points at prescribed intervals. When this access point 30 receives a packet from another access point 30 A, the routing control information processing unit 38 determined whether the packet is a data packet or a control packet containing control information. If the packet is a data packet, the packet is temporarily stored in the backbone-system receiving buffer 31 a.

The determination between data and control information may be performed at the determination unit (not shown in FIG. 5 ) provided in the backbone-system transmitting/receiving unit 35 , as illustrated in FIG. 2 . In this case, if the packet is determined to be a data packet, the data item is supplied to the backbone-system receiving buffer 31 a ; if determined to be control information, the packet is supplied to the routing control information processing unit 38 .

Upon receiving the control packet, the routing control information processing unit 38 estimates a node cost according to the amount of data accumulated in the receiving buffer 31 (backbone-system receiving buffer 31 a ) and a link cost according to the radio condition of the link between this access point 30 and the previous node (AP). A monitoring unit illustrated in FIG. 2 may be provided in the access point 30 so as to determine the amount of data accumulated in the receiving buffer 31 .

The routing control information processing unit 38 determines a cost value by, for example, multiplying the node cost by the link cost, or alternatively, adding the node cost to the link cost. The determined cost value is added to the current cost value contained in the control information to produce a total cost value required from the source node (AP) that generated the routing control packet to this node 30 . The produced total cost value is compared with another total cost value produced based on the routing control packet generated by the source node and received from another route to select the lower value. The selected total cost value is written in the routing control packet.

FIG. 6 illustrates an example of the above-described cost estimation. In this example, AP 1 is the source node that originally generates a routing control packet, and it transmits the routing control packet with cost value of zero. The routing control packet is then received at AP 2 and AP 5 . Access point AP 2 determines cost C 2 by adding the link cost C AP1-2 representing the radio condition (for example, the transmission rate) between AP 1 and AP 2 to the node cost C AP2 representing the traffic (for example, the amount of data currently accumulated in the receiving buffer). The AP 2 writes the cost value C 2 in the routing control packet and transmits the packet to adjacent access points.

Access point AP 3 received the routing control packet from AP 2 , and determines the total cost C 3 from AP 1 to AP 3 by adding the link cost C AP2-3 between this node and the previous node AP 2 and the node cost C AP3 representing the node traffic at AP 3 . The routing control packet containing the cost value C 3 is transmitted to adjacent access points. Access point AP 4 that has received the routing control packet performs the same process to determine the total cost C 4 .

On the other hand, the access point AP 4 also receives a routing control packet via another route through access points AP 5 and AP 6 , which nodes receive the routing control packet from the source node AP 1 and perform the above-described operations. The access point AP 4 determines the total cost value C 4 ′ for this route.

The routing control information processing unit 38 compares the total cost value C 4 ′ with the previously determined total cost value C 4 . If the newly determined total cost value C 4 ′ is smaller than the other cost value C 4 , the routing control information processing unit 38 rewrites the currently determined total cost value C 4 ′ over the previously determined total cost value C 4 in the routing control packet. The access point AP 4 then transmits the routing control packet to adjacent access points (not shown), while it records the previous node AP 6 as the next hop for the path leading to the access point A 1 according to the selected route. The next hop on the route reaching AP 1 is reported to the routing control unit 34 . This information is used when the access point AP 4 receives a packet addressed to AP 1 or the user terminal located under AP 1 .

Although comparison of the total cost values for different routes, rewriting of the total cost value, and recording of the next hop are explained in conjunction with AP 4 , the same operations are performed at each of the access points AP 2 , AP 3 , AP 5 , and AP 6 to select a route with the lowest cost in a self-directed manner. All the nodes (or access points) cooperate with each other to perform the above-described operations in a self-directed manner to build up the optimum route taking into account both the traffic at each node and the radio conditions of the wireless links.

When relaying packets, the routing control unit 34 designates the next hop, to which the data items temporarily stored in the receiving buffer 31 are to be transmitted, according to the lowest-cost route, and the data items (packets) are transmitted via the backbone-system transmitting/receiving unit 35 .

Although in FIG. 5 two radio interfaces (backbone-system transmitting/receiving unit 35 and the access-system transmitting/receiving unit 36 ) and two buffers (backbone-system receiving buffer 31 a and access-system receiving buffer 31 b ) are used, the above-described routing control can be realized using a single radio interface and a single buffer.

The routing control based on both the node cost and the link cost may be combined with the transmission priority control using the independently provided backbone-system receiving buffer and access-system receiving buffer illustrated in FIG. 2 and FIG. 3 . By combining these two features, prevention of workload concentration and efficient use of the network resources can be realized more effectively.

This patent application is based on and claims the benefit of the earlier filing dates of Japanese Patent Application No. 2004-119288 filed Apr. 14, 2004, the entire contents of which are hereby incorporated by reference.