Wireless routing selection system and method

Abstract

A technique involves untethered access points (UAPs) that can broadcast estimated transmission time (ETT) that represents an estimated time it would take for a packet to be transmitted from the first UAP to an AP that is wire coupled to a network. The proposed system can offer, among other advantages, accurate ETT values for use by UAPs of a wireless network.

Description

BACKGROUND

Next hop selection in a wireless protocol is made by selecting a least cost hop. Historically, cost has been determined by hop count, signal strength, error rate, utilization, and other factors. One technique for wireless routing selection involves defining cost based on expected transmission time (ETT) for some link (ETTI).

For example, link cost may be determined by measuring the transmission time to send a 1 Mbps stream of packets across the link and measuring its transmission time for some number of bytes. An algorithm may measure for each available bandwidth across the link, and the transmission time is defined as the time from when the packet is scheduled (specifically, sent to the radio) and the time that an acknowledgement is received.

The improvement of algorithms for next hop selection are the subject of research. Any improvements may have significant repercussions on the relevant technologies. Accordingly, any improvement in next hop selection would be advantageous.

These are but a subset of the problems and issues associated with wireless routing selection, and are intended to characterize weaknesses in the prior art by way of example. The foregoing examples of the related art and limitations related therewith are intended to be illustrative and not exclusive. Other limitations of the related art will become apparent to those of skill in the art upon a reading of the specification and a study of the drawings.

SUMMARY

The following embodiments and aspects thereof are described and illustrated in conjunction with systems, tools, and methods that are meant to be exemplary and illustrative, not limiting in scope. In various embodiments, one or more of the above-described problems have been reduced or eliminated, while other embodiments are directed to other improvements.

A wireless network system is typically coupled to a wired network at some point. Such a point is sometimes referred to as an access point (AP). A plurality of untethered APs (UAPs) may be coupled to one another, and eventually to the AP, to allow a wireless network to grow to practically any size. However, as the network grows in size using UAPs, it becomes more difficult to figure out a best path from a mobile station, through the UAPs to the AP in an optimal fashion.

Advantageously, UAPs can broadcast estimated transmission time (ETT) that represents an estimated time it would take for a packet to be transmitted from the first UAP to the AP. Thus, a UAP that is right next to the AP should be able to give a low ETT to the AP. As the advertised ETTs percolate through the wireless network, UAPs can eventually settle on optimal paths to the AP. The better the estimate, the more likely the optimally chosen paths are actually optimal.

The proposed system can offer, among other advantages, accurate ETT values for use by UAPs of a wireless network. This and other advantages of the techniques described herein will become apparent to those skilled in the art upon a reading of the following descriptions and a study of the several figures of the drawings.

BRIEF DESCRIPTION OF THE DRAWINGS

Embodiments of the invention are illustrated in the figures. However, the embodiments and figures are illustrative rather than limiting; they provide examples of the invention.

FIG. 1 depicts an example of a rate aware wireless system.

FIG. 2 depicts an example of a weighted graph of source, next hop, and destination nodes.

FIG. 3 depicts an example of a system in which an ETTp calculation includes time spent on an output queue.

FIG. 4 depicts a graph that provides a conceptual depiction of queue latency.

FIG. 5 depicts an example of a wireless network system that includes a plurality of untethered APs (UAPs).

FIG. 6 depicts a flowchart of an example of a method for selecting a next hop.

FIG. 7 depicts a flowchart of an example of a method for measuring ETTl to a node.

FIG. 8 depicts a flowchart of an example of a method for advertising an ETTp.

FIG. 9 depicts a flowchart of an example of a method for calculating NTT.

DETAILED DESCRIPTION

In the following description, several specific details are presented to provide a thorough understanding of embodiments of the invention. One skilled in the relevant art will recognize, however, that the invention can be practiced without one or more of the specific details, or in combination with other components, etc. In other instances, well-known implementations or operations are not shown or described in detail to avoid obscuring aspects of various embodiments, of the invention.

FIG. 1 depicts an example of a rate aware wireless system 100. In the example of FIG. 1, the system 100 includes a node 110, a node 120, and a node 130. For illustrative purposes, the node 110 and the node 130 are currently linked via active link 112, while the node 120 and the node 130 are not currently linked, as represented by the candidate link 122. In an embodiment, the candidate link 122 is periodically measured to determine if it is a better route than the active link 112. Optionally, if the node 130 is a next hop from a source node to a destination node, the node 130 may be linked to another node (not shown) through a next hop link 132.

In the example of FIG. 1, the node 110 advertises an estimated transmission time (ETT) for the path (ETTp) to a destination. ETTp 114 is the sum of ETT for each link (ETTl) from the source (e.g., the node 110) to the destination (not shown). ETTp 124 is the sum of ETTl from the source (e.g., the node 120) to the destination (not shown). Optionally, the node 130 advertises an ETTp 134 that is the ETTp from the node 130 to the destination (passing through either the node 110 or the node 120). ETTp 134 is optional because it will only exist if the node 130 is a next hop node.

FIG. 2 depicts an example of a weighted graph 200 of source, next hop, and destination nodes. The weights of the edges in the graph 200 are ETTl between two nodes of the graph 200. ETTp is the sum of ETTl from a source node 202 to a destination node 206. Typically, there are multiple next hop nodes 204-1 to 204-N (referred to collectively as nodes 204) between the source node 202 and the destination node 206, though it is possible to have none. As is shown in FIG. 2, the ETTl from the source node 202 to the node 204-1 has an ETTl₀. In general each of the nodes 204 has an ETTlx to the next hop, where x=the ordinal position of the current node. For example, the ETTl₁ is the ETTl from the node 204-1 to the node 204-2. As another example, the ETTl_N is the ETTl from the node 204-N to the destination node 206.

In some embodiments, the ETTp calculation is for the time a packet is sent from a radio until the time an acknowledgement is received. This, however, does not include time spent on a queue waiting for the radio to become available. Advantageously, by including the time spent on the queue, the ETTp calculation can take into consideration the real time it takes to transmit a packet based on load and utilization.

FIG. 3 depicts an example of a system 300 in which an ETTp calculation includes time spent on an output queue. In the example of FIG. 3, the system 300 includes a wireless device 302, an access point (AP) 304, and an AP 306, a wireless switch 308, and a wired network 310. It may be noted that the AP 304 is depicted as an untethered AP. In an embodiment, any number of untethered APs could be coupled together to reach the tethered AP 306.

In the example of FIG. 3, the wireless device 302 includes a queue 312, with packets 314-1 to 314-N enqueued thereon. The packet 314-1 is presumably a first packet of a stream of packets tha the wireless device 302 is trying to send to the AP 304. However, the AP 304 may not be available, which results in the packet being enqueued in the queue 312, as shown. The packet 314-N is the last packet to be enqueued prior to the packet 314-1 finally being sent to the AP 304. Thus, the example of FIG. 3 illustrates the queue 312 just before the packet 314-1 is sent to the AP 304 (and dequeued). The time spent waiting may be referred to as radio availability latency because it measures the time it takes for a radio (at the AP 304, in this case) to become available.

The AP 304 has a comparable queue 316, which is coupled to an ETT engine 318. The wireless device 302 may or may not have an ETT engine to determine how long a packet is enqueued on the queue 312, but in the example of FIG. 1, no such engine is present at the wireless device 302. The queue 316 functions in a manner quite similar to that described with reference to the queue 312. At the AP 304, however, the ETT engine 318 actually measures the amount of time a packet is enqueued. This radio availability latency can be added to an advertised ETTp, as described later with reference to FIG. 1, to give a more accurate measure of ETT for a packet.

Advantageously, ETT can be used by a next hop selector to decide upon an optimal next hop. In an embodiment, each AP includes a next hop selector.

FIG. 4 depicts a graph 400 that provides a conceptual depiction of queue latency. In the example of FIG. 4, the graph 400 includes (for illustrative purposes) a flat, or static, link rate 402 and a data rate 404 that increases over time. Where the link rate 402 is greater than the data rate 404, the link is under-utilized, as shown by the shaded link underutilization portion 406 of the graph 400. The link saturation point 408 is at a time where the link rate 402 and the data rate 404 are the same. At the link saturation point 408, the link is fully utilized. Where the link rate 402 is less than the data rate 404, the link is congested, as shown by the shaded link congestion portion 410 of the graph 410. When the link is congested, packets will arrive at an output queue, such as the queue 316 (FIG. 3) at a rate that is greater than the rate at which the packets are dequeued (and transmitted). Thus, the time spent waiting on the queue will grow as the link grows more congested. Advantageously, an ETT engine, such as the ETT engine 318 (FIG. 3) can measure this time spent waiting and incorporate the measurement into an ETT calculation.

From “A Radio Aware Routing Protocol for Wireless Mesh Networks” by Kulkarni et al. defines cost based on ETTl, and how ETTl can be aggregated to determine ETTp. However, the algorithm used by Kulkarni et al. can be improved in some specific cases. For example, the choice of 1 Mbps load rate for link cost calculation is arbitrary and may be significantly off. In an embodiment, expected load rate (ELR) is used instead. ELR is the load that a link would be subject to if it was selected as a next-hop.

Referring once again to the example of FIG. 1, an ELR 10-30136 and an ELR 30-10138 are associated with the active link 112. The ELR 10-30136 is intended to illustrate ELR from the node 110 to the node 130 and the ELR 30-10138 is intended to illustrate ELR from the node 130 to the node 110. In an embodiment, the ETT of a link will vary greatly depending on how much traffic is inserted into it. The more traffic you insert into a link, the higher the probability for collisions on the link. Accordingly, the ELR 10-30136 is calculated dynamically based on current load conditions of the active link 112 from the node 110 to the node 130, and the ELR 30-10138 is calculated dynamically based on current load conditions of the active link 112 from the node 130 to the node 110. The calculated ELR may be averaged in an exponentially decaying fashion to allow route selection stabilization.

In the example of FIG. 1, conceptually, the node 130 is trying to select the least cost link to some destination reachable through both the node 110 and the node 120. As shown in the system 100, the active link 112 has an ELR 10-30136 and an ELR 30-10138. The ELR 10-30136 and the ELR 30-10138 can be used to respectively calculate an effective data rate (EDR) 10-30116 and an EDR 30-10118.

EDR is the rate determined by a rate selection algorithm. In general, the rate selection algorithm should meet the following goals: 1) To the extent possible, the selected rate should produce optimal throughput of packets transmitted to a client. This is not necessarily the same thing as minimizing retries. For instance, retransmitting one time a large packet at 54 Mbps may result in better throughput than transmitting the same large packet at a 1 Mbps with no retries. 2) To the extent possible, the algorithm should be computationally light. That is, it should not consume a lot of CPU time to determine a rate to use.

An example of a rate selection algorithm is as follows (though any applicable known or convenient rate selection algorithm could be used): The rate selection algorithm seeks to minimize retransmissions. For each client it maintains a ‘best rate’ value. The rate selection algorithm is a control system that lowers the best rate when the rate of retransmissions exceeds 50% and raises the best rate when the rate of retransmissions is less than 50%. For each transmitted packet, there are one of three possible outcomes. 1) The packet is successfully transmitted with no retransmissions, 2) the packet is successfully transmitted with one or more retransmissions, 3) the packet transmission is unsuccessful after all retransmission attempts.

For each client, a counter is maintained. When a packet is successfully transmitted with no retransmissions, this counter is incremented by 3. When a packet is successfully transmitted but with retransmissions, the counter is decremented by 6. When a packet is not successfully transmitted, the counter is not changed. When the counter reached a value of −50, then the next lower rate is made the best rate. When the counter reaches a value of 100, the next higher rate is used as the best rate; however, the best rate is not increased if it has been increased in the past 60 seconds. This prevents the best rate from increasing too fast.

For each packet, transmissions are attempted using up to four rates.

• • The best rate is tried 1 time. This is the initial transmission attempt, not a retransmission.
  • The next best rate is tried for configured number of retransmissions minus 2. For example, the default value for the retry count is 5, and so by default the next best rate is tried 3 times.
  • The next lower rate is tried 1 time.
  • The lowest rate supported by the radio is tried 1 time.

This rate fall back schedule has the following properties. 1) If the best rate is successful, then there are no retries and the client's counter is increased. 2) If the best rate fails, then the next lower rate is used multiple times. The range of the next best rate is better than the best rate, and so the next best rate has a higher probability of success. The client's counter will be decremented in this case to reflect that the best rate was unsuccessful. 3) The radio's lowest rate has the best range, and so if it fails, then the client is not reachable or the failure is due to factors not related to distance. In this case, the client's counter is unchanged because the failure is not related to rate.

If the EDR is actually determined ELR, the algorithm further reduces the bandwidth required to compute ETTl, since the EDR need not be calculated through synthesized load. Notably, as shown in FIG. 1, the EDR 20-30126 uses the ELR 10-30136, and the EDR 30-20128 uses the ELR 30-10138. Accordingly, for the candidate link 122 as well, a synthesized load is not used. Advantageously, in both cases, ELR is calculated based on existing traffic.

It should be noted that sensing all data rates is less efficient than using the techniques described herein. Advantageously, by using EDR, all possible rates need not be tested, making this technique more efficient. Moreover, selected rates may not be the rate actually selected by a radio transmission module. For example, if data rate selection does not yield an answer that matches an algorithm such as Kulkarni's, the actual ETTl will be different than the expected ETTl and the algorithm will make suboptimal decisions. So using EDR can lead to performance improvements as well.

In an embodiment, the ETTp calculation can be improved by considering the amount of time a packet spends being processed in intermediate nodes. This is the time it takes to receive a packet on some interface and queue it on its egress interface. This time is referred to as node transit time (NTT). Therefore, in a non-limiting embodiment, ETTp=ETTI+ETTp_nh +NTT, where ETTI is the link between a node and a next hop node, ETTp_nh is the ETTp advertised by the next hop node (e.g., the best advertised ETTp of potential next hop nodes), and NTT is the time a packet spends transiting a node. As was previously described, the ETT calculations include the time a packet spends in a queue waiting for a radio to become available. Conceptually, the NTT is the time a packet spends in a node waiting to be enqueued.

The techniques described herein work best when there are relatively few interesting destinations. Advantageously, this is exactly the case in most IP network environments. Most hosts are trying to communicate to their next hop IP router, which is typically eventually accessed over a wired network. Hence, the techniques described herein help answer the question “how do I get to the wired network?” Only a single destination need be evaluated and only a single value to ELR needs to be maintained.

FIG. 5 depicts an example of a wireless network system 500 that includes a plurality of untethered APs (UAPs). In the example of FIG. 5, the system 500 includes a UAP 502, a UAP 504, a plurality of UAPs 506-1 to 506-N (referred to collectively as UAPs 506), and an AP 508. For illustrative purposes only, a path for wireless traffic from a station 510 to the AP 508 is depicted as a dashed line. Potential paths for wireless traffic from the station 510 to the AP 508 are depicted as dotted lines.

In the example of FIG. 5, wireless traffic from the station 510 is directed to an AP with which the station 510 has associated. Typically, though not always, the AP with which the station associates is the one that is closest to the station 510 (or the one that detects the highest RSSI from the station 510). In the example of FIG. 5, the closest station is presumed to be the UAP 502.

In the example of FIG. 5, presumably, at some stage it was determined that the best path from the station 510 to the AP 508 was from the USP 502 to the UAP 504 and finally to the AP 508. However, the system 500 continuously or occasionally measures ETT for various nodes, as was described above. Thus, it may be determined that a different path (through one of the UAPs 506) is better. It should be noted that, depending upon the implementation and/or embodiment, a tethered AP could be rejected as a next hop in favor of a UAP, followed by an eventual hop to some other AP. This would be the case if ETTp from the UAP was better than the ETTp directly to the tethered AP. Presumably, this would be unusual, but not impossible.

At the UAP 502, the goal is to send traffic to the least expensive AP that is wired to a network. By least expensive, what is intended is that a weighted graph with edges that are ETT between nodes, would yield the smallest result possible (or practical). This AP may or may not be the AP closest to the UAP 502. The UAP 502, for illustrative purposes, is illustrated as a large circle with various components. However, the UAP 504, the UAPs 506, and/or the AP 508 may have similar components (not shown).

In the example of FIG. 5, the UAP 502 includes an ingress interface 512, an ETTp engine 514, a next hop selector 516, and an egress interface 518. The ETTp engine 514 includes an ETTp_nh module 520, an NTT module 522, and an ETTl module 524. In operation, in a non-limiting embodiment, the UAP 504 and the UAPs 506 have broadcast advertised ETTp values that are associated with the path from the respective nodes to a destination, such as the wired network. The ETTp_nh module 520 receives each of the advertised ETTps.

Some time later (or concurrently) the station 510 sends packets to the UAP 502, which are received at the ingress interface 512. The NTT module 522 receives an indication, such as a first timestamp, that a first packet has been received. As much as is practical, it would probably be valuable to have the timestamp represent the exact time the first packet was received at the ingress queue 512, though an estimate may be used. At this point, the ETTp engine 514 knows only ETTp values for the UAP 504 and UAPs 506, but has no link information. It should be noted that in practice there will typically be link information as described later. Nevertheless, assuming for a moment that no link information is available, the ETTp engine 514 can provide the advertised ETTp values to the next hop selector 516, which picks an appropriate optimal path to the destination based on the advertised ETTp values. Specifically, the next hop selector 516 chooses the shortest (e.g., lowest weight) path to the destination.

The first packet is enqueued at the egress interface 518, as appropriate. It may be noted that the first packet may or may not need to be enqueued in a case where the relevant link is underutilized (or saturated but not congested). In any case, when the first packet is received at the egress interface 518, the NTT module 522 receives an indication, such as a second timestamp, that the first packet has been received at the egress interface 518. At this point, the NTT module 522, by comparing, for example, a first timestamp and a second timestamp, can calculate the amount of time that the first packet spent at the UAP 502. This information is useful for purposes that are described below.

The first packet is sent from the egress interface 518 to the UAP 504. For illustrative purposes, it is assumed that the UAP 504 is the next hop in an optimal path. In a non-limiting embodiment, the UAP 504 sends an acknowledgement, as soon as the first packet is received, that the first packet was received. The acknowledgement is received at an acknowledgement interface 526. It should be noted that the acknowledgement interface 526 may be part of a radio interface that includes the ingress interface 512 (or even the egress interface 518). In any case, the acknowledgement interface 526 provides the ETTl module 524 with an indication, such as a timestamp, that an acknowledgement was received from the next hop node. The ETTl module 524 uses the indication (e.g., second timestamp) that was generated when the first packet was enqueued on the egress interface 518 and the indication (e.g., third timestamp) that was generated upon receipt of the acknowledgement to provide an ETTl value.

At this point, the ETTp engine 514 has enough information to know ETTp from the UAP 502 to the destination. Specifically, ETTl+NTT+ETTp_nh=ETTp from the UAP 502 to the destination. This ETTp value can be provided to an ETTp broadcast engine 528. In the example of FIG. 5, the broadcast engine 528 is not providing any value to the station 510 (unless the station 510 includes a means for making use of the broadcast ETTp). However, the UAP 504, for example, may have a broadcast engine that functions similarly. Such an engine could be used to provide the advertised ETTp to the ETTp_nh module 520, as described previously.

FIG. 6 depicts a flowchart 600 of an example of a method for selecting a next hop. In the example of FIG. 6, the flowchart 600 starts at module 602 where ETTp is received from nodes that are within range. In an embodiment, the node at which a next hop is being selected listens for any node within range. In an alternative, the potential next hop nodes may be restricted in some manner.

In the example of FIG. 6, the flowchart 600 continues to module 604 where ETTl is measured to each node within range. Since the ETTl is an actual measurement (rather than a guess), the ETTl is a relatively accurate representation of actual link characteristics. Any applicable known or convenient technique may be used to measure ETTl. An example of a method for measuring ETTl to a node is described later with reference to FIG. 7.

In the example of FIG. 6, the flowchart 600 continues to module 606 where ETTl is added to ETTp from each node to arrive at a node-specific path metric, and to module 608 where a next hop is selected that is associated with a minimum of the node-specific path metrics. Notably, the lowest ETTp plus a corresponding ETTl is not necessarily lower than some other ETTp plus a corresponding ETTl.

FIG. 7 depicts a flowchart 700 of an example of a method for measuring ETTl to a node. In the example of FIG. 7, the flowchart 700 starts at module 702 where a packet is placed on an egress queue. Packets are placed on egress queues when they are ready to be transmitted to a next hop or destination.

In the example of FIG. 7, the flowchart 700 continues to modules 704 where a first timestamp is taken. The first timestamp represents the approximate time at which the packet was placed on the egress queue. The packets may be left on an egress queue for a relatively long time if they are enqueued at a faster rate than they are dequeued (and transmitted). Typically, if a packet remains in the egress queue for a relatively long period of time, a link between the current queue and the next hop or destination is congested.

In the example of FIG. 7, the flowchart 700 continues to module 706 where an acknowledgement is received that the packet was transmitted. The acknowledgement may be in the form of, by way of example but not limitation, an 802.11 ack. Other protocols may have other techniques or terminologies, but any applicable known or convenient means for acknowledging that the packet was received may be used, depending upon the implementation and/or embodiment.

In the example of FIG. 7, the flowchart 700 continues to module 708 where a second timestamp is taken. The second timestamp represents the approximate time at which the packet that was placed on the egress queue, plus the time to reach the next hop, plus the time to receive the acknowledgement (which is normally sent immediately upon receipt of the packet). Alternatively, the second timestamp could be placed in the acknowledgement such that the time to receive the acknowledgement is omitted.

In the example of FIG. 7, the flowchart 700 continues to module 710 where a difference between the first timestamp and the second timestamp is found. In a non-limiting embodiment, this entails calculating an exponentially decaying average of the difference. In any case, the value found may be used as an ETTl.

FIG. 8 depicts a flowchart 800 of an example of a method for advertising an ETTp. In the example of FIG. 8, the flowchart 800 starts at module 802 where an advertised ETTp is calculated. ETTp is calculated by selecting an advertised ETTp from some other node and adding local NTT. NTT may be, by way of example but not limitation, an exponentially weighted average of the time it takes to transmit a packet from an ingress to an egress queue in a node. An example of a method for calculating NTT is described later with reference to FIG. 9.

In the example of FIG. 8, the flowchart 800 continues to module 804 where the advertised ETTp is broadcast. In an alternative embodiment, the ETTp may be multicast to a subset of nodes within broadcast range. Any nodes within range may use the advertised ETTp when selecting a next hop, if applicable.

FIG. 9 depicts a flowchart 900 of an example of a method for calculating NTT. In the example of FIG. 9, the flowchart 900 starts at module 902 with receiving a packet on an ingress interface. The packet may be received from a wireless station, such as a mobile device or UAP.

In the example of FIG. 9, the flowchart 900 continues to module 904 where a first timestamp is taken. The first timestamp represents the point in time when the packet is first received at the node.

In the example of FIG. 9, the flowchart 900 continues to module 906 where the packet is forwarded to an appropriate egress interface. Techniques for forwarding packets to egress interfaces are well known in the relevant art, and are not described herein. It is assumed that some applicable known or convenient technique is used.

In the example of FIG. 9, the flowchart 900 continues to module 908 where a second timestamp is taken. The second timestamp represents the point in time when the packet has been enqueued for sending to a next hop or destination.

In the example of FIG. 9, the flowchart 900 continues to module 910 where a difference between the first timestamp and the second timestamp is found. In a non-limiting embodiment, an exponentially decaying average is used. In an y case, the derived value may be used as the local NTT.

As used herein, access point (AP) refers to receiving points for any known or convenient wireless access technology. Specifically, the term AP is not intended to be limited to 802.11 APs.

Some portions of the detailed description are presented in terms of algorithms and symbolic representations of operations on data bits within a computer memory. These algorithmic descriptions and representations are the means used by those skilled in the data processing arts to most effectively convey the substance of their work to others skilled in the art. An algorithm is here, and generally, conceived to be a self-consistent sequence of operations leading to a desired result. The operations are those requiring physical manipulations of physical quantities. Usually, though not necessarily, these quantities take the form of electrical or magnetic signals capable of being stored, transferred, combined, compared, and otherwise manipulated. It has proven convenient at times, principally for reasons of common usage, to refer to these signals as bits, values, elements, symbols, characters, terms, numbers, or the like.

It should be borne in mind, however, that all of these and similar terms are to be associated with the appropriate physical quantities and are merely convenient labels applied to these quantities. Unless specifically stated otherwise as apparent from the following discussion, it is appreciated that throughout the description, discussions utilizing terms such as “processing” or “computing” or “calculating” or “determining” or “displaying” or the like, refer to the action and processes of a computer system, or similar electronic computing device, that manipulates and transforms data represented as physical (electronic) quantities within the computer system's registers and memories into other data similarly represented as physical quantities within the computer system memories or registers or other such information storage, transmission or display devices.

The algorithms and techniques described herein also relate to apparatus for performing the algorithms and techniques. This apparatus may be specially constructed for the required purposes, or it may comprise a general purpose computer selectively activated or reconfigured by a computer program stored in the computer. Such a computer program may be stored in a computer readable storage medium, such as, but is not limited to, read-only memories (ROMs), random access memories (RAMs), EPROMs, EEPROMs, magnetic or optical cards, any type of disk including floppy disks, optical disks, CD-ROMs, and magnetic-optical disks, or any type of media suitable for storing electronic instructions, and each coupled to a computer system bus.

As used herein, the term “embodiment” means an embodiment that serves to illustrate by way of example but not limitation.

It will be appreciated to those skilled in the art that the preceding examples and embodiments are exemplary and not limiting to the scope of the present invention. It is intended that all permutations, enhancements, equivalents, and improvements thereto that are apparent to those skilled in the art upon a reading of the specification and a study of the drawings are included within the true spirit and scope of the present invention. It is therefore intended that the following appended claims include all such modifications, permutations and equivalents as fall within the true spirit and scope of the present invention.

Claims

1. A method comprising: receiving, at a current wireless node, an estimated transmission time (ETT) from each of a plurality of other wireless nodes that are within a range of the current wireless node, wherein each said ETT is a next-hop-to-destination-path ETT (ETTp) associated with one of said other wireless nodes, wherein each said ETTp is a function of a link ETT (ETTI), another ETTp and a node transition time (NTT) of one of said other wireless nodes;at the current wireless node, measuring for each of said other wireless nodes an ETTI between the current wireless node and a corresponding one of said other wireless nodes;at the current wireless node, adding for each of said other wireless nodes an ETTI to a corresponding ETTp to determine a node-specific path metric for a corresponding one of said other wireless nodes;at the current wireless node, selecting a next hop based on the determined node-specific path metrics;at the current wireless node, calculating an advertised ETTp for the current wireless node, based on at least one of the measured ETTI's, a corresponding at least one of the received ETTp's and a calculated NTT for the current wireless node; andbroadcasting the advertised ETTp from the current wireless node to a second wireless node.

2. The method of claim 1, wherein measuring an ETTI comprises: placing a packet on an egress queue;taking a first timestamp;receiving acknowledgement that the packet was transmitted;taking a second timestamp;finding the difference between the first timestamp and the second timestamp.

3. The method of claim 2, wherein finding the difference between the first timestamp and the second timestamp includes taking an exponentially decaying average.

4. The method of claim 1, wherein the calculating the advertised ETTp comprises: receiving a packet on an ingress interface;taking a first timestamp;forwarding the packet to an appropriate egress interface;taking a second timestamp;finding the difference between the first timestamp and the second timestamp.

5. The method of claim 4, wherein finding the difference between the first timestamp and the second timestamp includes taking an exponentially decaying average.