KR20100042729A - Method for executing a handoff in wireless ad hoc network - Google Patents

Abstract

PURPOSE: A method for executing a handoff in a wireless ad hoc network is provided to perform an adaptive handoff between a mobile ad hoc network and an Internet network. CONSTITUTION: A current traffic load is compared with a previous traffic load(S304). If the previous traffic load is larger than the current traffic load, the difference between previous traffic load the current traffic load is compared with a second setup value(S314,S316). If the difference between the previous traffic load and the current traffic load is larger than the second setup value, a hop number weight value of cost function is increased(S318). The surplus bandwidth weight value is reduced.

Description

Method for executing a handoff in wireless ad hoc network}

The present invention relates to a method of performing a handoff in an ad hoc network, and more particularly, to a method of enabling a node of a mobile ad hoc network to adaptively handoff between a mobile ad hoc network and the Internet.

With the development of electronic and communication technologies, users can send and receive data anytime, anywhere through the network. As described above, a network for transmitting and receiving data includes a wired network that directly connects electronic devices using a wire to form a network, and a wireless network that forms a network without using a wire.

The wireless network includes all types of computer networks formed wirelessly. A wireless network can communicate without using wires between network nodes. The wireless network may be classified into a wireless personal area network (WPAN), a wireless local area network (WLAN), a wireless metropolitan area network (WMAN), a wireless wide area network (WWAN), and the like, according to a range in which the network is formed.

The wireless network may be connected to a fixed wired network, or may be a network consisting only of a mobile host without being connected to a fixed wired network. A network composed only of mobile hosts, not connected to the wired network, is called a mobile ad-hoc network (or a mobile ad-hoc network). Each mobile host (or mobile node) in a mobile ad hoc network functions as one router as well as a host and may have multiple paths to other nodes.

MANET is a wireless network that uses multihop communication between mobile nodes located in the same area without using an infrastructure such as a base station, and can configure a network at low cost. As the demand for the Internet service increases, research is being actively conducted to provide an Internet access service to a mobile node of a MANET by connecting a MANET composed of mobile nodes only to the Internet.

1 is a diagram schematically illustrating a network structure for connecting a MANET to the Internet. An Internet gateway (IGW) router is located between the MANET and the Internet network. As shown in FIG. 1, a network composed of a MANET, an Internet gateway router, an Internet network, and the like is called a Connected-MANET or a Hybrid-MANET. In addition, a wireless access network using multiple hops to provide Internet access to a mobile node in Connected-MANET is referred to as a multihop radio access network (MRAN).

In MANET, mobile nodes connect to the Internet through the Internet gateway router. The mobility of the nodes allows the nodes to perform handoff, which is a process of changing the Internet gateway router in communication. In a cellular mobile communication network, the signal reception strength of a signal is measured to perform a handoff to a base station having the highest strength. However, in a MANET in which a mobile node is connected to an Internet gateway router through multiple hops, it is difficult to perform a handoff based on the reception strength of a signal from the Internet gateway router. Handoff should be performed based on the overall quality of the multipath between the mobile node and the Internet gateway router.

The present invention provides a method in which a mobile node wishing to access the Internet in a mobile ad hoc network can adaptively handoff between the mobile ad hoc network and the Internet network.

The handoff method according to the present invention comprises the steps of: setting an initial value of a hop number weight value and a surplus bandwidth weight value of a cost function at a node, measuring a current traffic load on a network at the node, the current traffic load and previous traffic Comparing the load with the current traffic load, and comparing the difference between the current traffic load and the previous traffic load with a first set value, wherein the difference between the current traffic load and the previous traffic load is greater than the first set value. If greater, decreasing the hop count weight value of the cost function and increasing the excess bandwidth weight value; comparing the current traffic load with a previous traffic load and if the previous traffic load is large, Comparing the difference with a second set point, and the previous traffic load and the current If the difference in the re-traffic load is greater than the second set value, increasing the hop number weight value of the cost function and decreasing the excess bandwidth weight value.

In another aspect, the handoff method includes measuring a current traffic load on a network at a node, and using a weight table according to the current traffic load and the traffic load, a hop number weight value and a surplus bandwidth weight value of a cost function. Determining a value, and handing off to the Internet gateway router with the lowest cost using the cost function from which the weight value is determined.

In another aspect, the handoff method includes setting an initial value of a hop number weight value and a surplus bandwidth weight value of a cost function at a node, measuring a message propagation delay time on a network at the node, and the current message. Comparing the delivery delay time with the previous message delivery delay time, if the current message delivery delay time is large, comparing a difference between the current message delivery delay time and the previous message delivery delay time with a first set value; If the difference between the delay time and the previous message propagation delay time is greater than the first set value, decreasing the hop number weight value of the cost function and increasing the surplus bandwidth weight value, the current message propagation delay time and the previous message propagation Compare the delay time, if the previous message delivery delay time is large Comparing the difference between the previous message delivery delay time and the current message delivery delay time with a second setting value, and if the difference between the previous message delivery time and the current message delivery delay time is greater than the second setting value, Increasing the hop number weight value of the cost function and decreasing the excess bandwidth weight value.

In another aspect, the handoff method includes measuring a message propagation delay time on a network at a node, and using a weight table based on the message propagation delay time and the message propagation delay time, Determining a surplus bandwidth weight value, and handing off to the Internet gateway router with the lowest cost using the cost function from which the weight value is determined.

In another aspect, the handoff method includes setting an initial value of a hop number weight value and a surplus bandwidth weight value of a cost function at a node, measuring a surplus bandwidth on a network at the node, and comparing the current surplus bandwidth with the current surplus bandwidth. Comparing a previous surplus bandwidth and comparing the current surplus bandwidth with a first set value, wherein the difference between the current surplus bandwidth and the previous surplus bandwidth is equal to the first set value. Increasing a hop number weight value of the cost function and decreasing a surplus bandwidth weight value, comparing the current surplus bandwidth with a previous surplus bandwidth, and if the previous surplus bandwidth is large, the previous surplus bandwidth and the current surplus Comparing the difference in bandwidth with a second set value, and said previous excess bandwidth The current difference of the excess bandwidth is greater than the second set value, and a step to reduce the hop number weighting value of the cost function and to increase the excess bandwidth weight value.

In another aspect, the handoff method includes measuring a current surplus bandwidth on a network at a node, a hop number weight value and a surplus bandwidth weight value of a cost function using a weight table according to the current surplus bandwidth and the surplus bandwidth. And handing off to the Internet gateway router with the lowest cost using the cost function from which the weight value is determined.

According to the method for performing handoff in the ad hoc network of the present invention, it is possible to efficiently manage mobile nodes and to efficiently use resources of the network. The mobile nodes can adaptively perform handoff in consideration of the network situation. For example, an optimal internet gateway router capable of accessing the Internet may be adaptively selected in consideration of the number of hops and the surplus bandwidth, and handoff may be performed. Therefore, it is possible to prevent limited network resource waste and increase efficiency of the wireless network.

Each node can efficiently perform handoffs with less strain on the network and can provide users with accurate and seamless Internet services.

The objects, features and advantages of the present invention will become apparent from the following detailed description of embodiments taken in conjunction with the accompanying drawings. In addition, the terms used in the present invention was selected as a general term widely used as possible now, but in certain cases, the term is arbitrarily selected by the applicant, in which case the meaning is described in detail in the corresponding description of the invention, It is to be clear that the present invention is to be understood as the meaning of terms rather than names.

The operation of the method for performing handoff in the ad hoc network according to the present invention configured as described above will be described in detail with reference to the accompanying drawings.

In a multi-hop radio access network (MRAN), handoffs generated by movement of mobile nodes are classified into forced handoffs and optimized handoffs according to the conditions in which handoffs occur. Forced handoff is performed when the mobile node is disconnected from the currently communicating and communicating Internet gateway router. On the other hand, optimal handoff is performed when a better Internet gateway router is found, even if the connection with the currently communicating Internet gateway router is not broken.

In forced handoff and optimal handoff, the mobile node must discover a new Internet gateway router before handing off. In a multi-hop wireless access network, there are largely a proactive discovery method and a reactive discovery method for a mobile node to discover an internet gateway router. In the proactive discovery scheme, the Internet gateway router periodically broadcasts an Internet gateway router advertisement message. In the reactive discovery scheme, when the mobile node requests an advertisement message from the Internet gateway router, the Internet gateway router transmits the Internet gateway router advertisement message in response.

Forced handoff may operate in both a proactive router discovery technique in which the gateway router periodically broadcasts its own information and a reactive router discovery technique in which the gateway broadcasts information at the request of the terminal. On the other hand, optimal handoff can only work with proactive router discovery techniques.

When performing a forced handoff or an optimal handoff, the mobile node selects an optimal Internet gateway router among a plurality of Internet gateway routers found through the Internet gateway router discovery technique and performs a handoff. The number of hops may be used as a metric for selecting the optimal internet gateway router. That is, an Internet gateway router with a lower number of hops from a particular mobile node is considered to be able to provide better communication performance so that the mobile node can handoff to an Internet gateway router with the lowest number of hops from among the selectable Internet gateway routers. Will perform.

However, when selecting the Internet gateway router by the minimum number of hops as described above, there is a disadvantage that it is not possible to properly meet the quality of service of communication sessions having different delay and bandwidth service requirements. Therefore, as shown in Equation 1 below, the cost function considering the hop number and the minimum surplus bandwidth corresponding to the j th routing path from the gateway router i is defined and handoff is performed to the Internet gateway router and the routing path having the minimum cost function. Do it.

및

는 각각 홉에 대한 가중치와 최소 잉여 대역폭에 대한 가중치를 나타내며,

및

는 각각 인터넷 게이트웨이 라우터 i로부터 j번째 라우팅 경로에 해당하는 홉 수 및 최소 잉여 대역폭을 나타내며, 이는 인터넷 게이트웨이 라우터 및 경로 선택 메트릭에 해당한다. 이 때,

를 가정한다.In Equation 1

And

Represents the weight for hop and the weight for minimum surplus bandwidth, respectively.

And

Denotes the hop number and the minimum surplus bandwidth corresponding to the j th routing path from the Internet gateway router i, respectively, and correspond to the Internet gateway router and the route selection metric. At this time,

Assume

) 및 최소 잉여 대역폭(

)을 들었으나, 구현 예에 따라 해당 노드의 배터리 잔여량이나 패킷 전달률 등을 비용함수를 정의하는 메트릭으로 사용할 수 도 있다. 이들은 상기 수학식 1의 추가적인 메트릭이 될 수도 있고, 홉 수나 최소 잉여 대역폭을 대체하는 메트릭이 될 수도 있다. 배터리 잔여량이나 패킷 전달률은 최소 잉여 대역폭과 같이 값이 클수록 비용함수에 기여하는 값이 작아진다.In Equation 1, as an example of a metric defining a cost function, hop number (

) And minimum surplus bandwidth (

However, depending on the implementation, the remaining battery of the node or the packet forwarding rate may be used as a metric defining the cost function. These may be additional metrics of Equation 1 above, or may be metrics that replace hop count or minimum surplus bandwidth. The higher the value, such as the minimum remaining bandwidth or the battery delivery rate, the smaller the value that contributes to the cost function.

In the multi-hop handoff using the cost function as shown in Equation 1, the performance of the handoff technique is affected by the weight value, and the optimal weight has different values depending on the traffic load. For example, when the traffic load on the network is light, the weight (w1 or W _hop ) for the number of hops is greater than the weight (w2 or W _BW ) for the excess bandwidth. On the contrary, in a heavy traffic load environment, the weight w2 of the surplus bandwidth is greater than the weight w1 of the number of hops. And, if the traffic load is moderate, it has excellent performance in properly maintaining weights for the number of hops and the surplus bandwidth.

Therefore, in the method of performing the handoff using the cost function, the method of performing the handoff using a fixed weight value does not provide optimum performance in a dynamic network environment in which the traffic of the network changes. . Hereinafter, a method of providing excellent handoff by adaptively adjusting a weight of a cost function to network traffic will be described.

FIG. 2 is a diagram schematically showing traffic loads classified according to network conditions according to an embodiment of the present invention. In FIG. 2, the traffic load of the network is divided into three stages.

In a network environment where the traffic load is L _low or less, it is preferable to increase the weight w1 of the hop number, and in an environment where the traffic load is L _high or more, it is preferable to increase the weight weight (w2) of the bandwidth. In an environment where the traffic load has a value between L _low and L _high , it is preferable to set an appropriate hop number and bandwidth weight. The optimal L _low and L _high values can be different depending on the network environment. The L _low and L _high values may be set differently according to an implementation example based on measurement data or simulation results in an actual network.

There may be various ways to adjust the weight based on the traffic load, but the following description will be divided into two methods.

3 is a diagram schematically illustrating a first example of adjusting a weight based on a traffic load according to an embodiment of the present invention.

In the first method, the network traffic load is measured by setting the hop count weight and the surplus bandwidth weight as initial values. If the traffic has increased above the threshold, the current measured value increases above the threshold, increasing the weight of the surplus bandwidth and decreasing the weight of the hop count. On the other hand, if the current measured value decreases above the threshold than the previously measured value, the weight of the hop number is increased and the weight of the excess bandwidth is reduced. If the change between the previously measured value and the currently measured value is less than the threshold value, the weight value is maintained. The initial value of the hop count weight and the surplus bandwidth weight may be selected in advance according to the network traffic conditions. For example, an initial value may be set differently according to an embodiment, and as a simple embodiment, there may be a method of selecting 1/2 each.

The first example will be described with reference to FIG. 3. Each mobile node initially sets the hop number weight and the surplus bandwidth weight and changes the weight value according to the change of traffic. Each mobile node initially sets the hop number weight and the surplus bandwidth weight to w_hop_init and w_bw_int, respectively (S300). Each mobile node has an initial weight and uses it to calculate the cost function and perform handoff.

Each mobile node periodically measures the traffic load (L _curr ) on the network (S302). Each node compares the currently measured traffic load (L _curr ) with the previous traffic load (L _prev ) (or the traffic load measured when the weight was most recently changed) (S304), and the current traffic load (L _curr ). If is large, it is determined that the traffic is increased (S314), and if the previous traffic load (L _prev ) is large, it is determined that the traffic is reduced (S306).

If it is determined that the traffic is reduced (S306), the difference between the current traffic load and the previous traffic load (L _prev -L _curr ) and the threshold value (L _def1 ) is compared (S308). If the difference between the current traffic load and the previous traffic load (L _prev -L _curr ) is greater than the threshold (L _def1 ), the node increases the weight for the hop count by △ if the weight for the hop count is less than 1- △. (W1 = w1 + Δ), and if the weight for the hop number is greater than 1-Δ, the weight for the hop number is 1 (S310). If the weight of the surplus bandwidth is greater than Δ, the node reduces the weight of the surplus bandwidth by Δ (w2 = w2-Δ), and if the weight of the excess bandwidth is less than Δ, the weight of the surplus bandwidth is 0. (S310).

As a result of the determination of step S308, if the difference between the current traffic load and the previous traffic load (L _prev -L _curr ) is smaller than the threshold value L _def1 , the node has a weight w1 for the number of hops and a weight for the surplus bandwidth ( w2) is maintained at the previous value (S312).

If it is determined that the traffic has increased (S314), the difference between the current traffic load and the previous traffic load (L _curr -L _prev ) and the threshold value (L _def2 ) is compared (S316). If the difference between the current traffic load and the previous traffic load (L _curr -L _prev ) is greater than the threshold (L _def2 ), the node decreases the weight for the number of hops by △ if the weight for the number of hops is greater than △ (w1 = w1-Δ), if the weight for the number of hops is less than △, the weight for the number of hops is set to 0 (S318). If the weight of the surplus bandwidth is less than 1-Δ, the node increases the weight of the surplus bandwidth by Δ (w2 = w2 + Δ), and if the weight of the surplus bandwidth is greater than 1-Δ, The weight is set to 1 (S318).

As a result of the determination in step S316, if the difference between the current traffic load and the previous traffic load (L _curr -L _prev ) is less than the threshold value (L _def2 ), the node has weight (w1) for the number of hops and weight for the surplus bandwidth ( w2) is maintained at the previous value (S312).

After changing the weight, each node stores the currently measured traffic load (L _curr ) in the previous traffic load (L _prev ) value (S320) for use in the next weight adjustment procedure. If the weight value is maintained as is through step S312, the currently measured traffic load L _curr is not stored in the previous traffic load L _prev value. Steps S302 to S320 are periodically performed at each node, and each node may adaptively adjust a weight value.

4 and 5 are diagrams schematically illustrating a second example of adjusting a weight based on a traffic load according to an embodiment of the present invention. In the second example, the weight value of the cost function is defined as a table in advance according to the traffic load interval classified in advance, and the weight value is adjusted accordingly.

4 is a diagram schematically showing a detailed classification of traffic load according to a network state according to an embodiment of the present invention. In FIG. 4, traffic loads between L _low and L _high are defined in N-1 steps. Each node defines a weight value of a cost function as a table in advance according to the traffic load interval classified in advance as shown in FIG. 4, and adjusts the weight value based on the weight value. That is, when the traffic load is small, the weight of the number of hops is greater than the weight of the excess bandwidth, and when the traffic load is large, the weight of the excess bandwidth is made larger than the weight of the number of hops to improve performance. When the traffic load is between L _low and L _high , the traffic load is divided into N-1 stages, and the weight of the cost function is appropriately adjusted according to the traffic load.

5 is a diagram schematically showing a weight table according to traffic load according to an embodiment of the present invention.

Each node may determine the weight w1 of the hop number of the cost function and the weight w2 of the surplus bandwidth using the weight table as shown in FIG. 5 and the currently measured traffic load value. In the weight table of FIG. 5, each weight value may be set in advance through actual measurement or simulation, and may vary according to implementation examples. The value of function f can be set appropriately based on network traffic and performance. For example, set f _i (·) = i / N (stepped function) regardless of network load, or f _i (·) = (e ⁱ -1) / (e ^N -1) (exponential function) ), f _i (·) = log (i + 1) / log (N + 1) (log function).

Each mobile node periodically measures the traffic load, and calculates a cost function by selecting weights defined in the weight table according to the measured traffic load. Each node performs a handoff to the least expensive internet gateway router using the calculated cost function value. This process can be repeated periodically.

In each of the above examples, a method of performing a handoff under the assumption that each mobile node can measure the load of traffic of the entire network is described. However, if it is difficult for each mobile node to directly measure the traffic load of the entire network, the indirect measurement allows the handoff to be performed by inferring the traffic load of the entire network.

Hereinafter, a method of inferring total network traffic by using an internet gateway advertisement message delay from an Internet gateway router and a method of inferring total network traffic using a minimum surplus bandwidth will be described. Describe how to perform handoff adaptively.

In the case of inferring the whole network traffic using the gateway advertisement message delay from the Internet gateway router, each gateway router broadcasts the Internet gateway router advertisement message to the mobile node at regular intervals. At this time, each Internet gateway router broadcasts by inputting a time stamp value in the advertisement message, which is time information when the advertisement message is broadcast. Accordingly, each mobile node can measure the time delay that it takes for the advertisement message to be delivered to itself by comparing the timestamp value extracted from the received advertisement message with the time it received the advertisement message. In general, when the traffic load on the network increases, the time taken to deliver a packet increases, so that the traffic load can be indirectly measured based on the delay of the advertisement message.

As described above, a process of adaptive handoff using the entire network traffic inferred using the gateway advertisement message delay is described below.

6 is a diagram schematically illustrating a third example of adjusting a weight based on a traffic load according to an embodiment of the present invention. 6 is a method of inferring traffic load by using a gateway advertisement message delay from an internet gateway router and changing a weight value of a cost function according to traffic change.

Each mobile node initially sets the hop number weight and the surplus bandwidth weight and changes the weight value according to the change of traffic. Each mobile node initially sets the hop number weight and the surplus bandwidth weight to w_hop_init and w_bw_int, respectively (S600). Each mobile node has an initial weight and uses it to calculate the cost function and perform handoff.

Each mobile node periodically measures the message propagation delay on the network (S602). The difference T _node -T _gw between the time information T _gw and the current time value T _node included in the advertisement message broadcast and received from the Internet gateway router is obtained and defined as a delay time T _{delay_curr} .

Each node compares the currently measured propagation delay time T _{delay_curr} with the previous propagation delay time T _{delay_prev} (or the propagation delay time measured when the weight was most recently changed) (S604). If (T _{delay_curr} ) is large, it is determined that the traffic is increased (S614), and if the previous transmission delay time (T _{delay_prev} ) is large, it is determined that the traffic is reduced (S606).

If it is determined that the traffic is reduced (S606), the difference between the current propagation delay time and the previous propagation delay time (T _{delay_prev} -T _{delay_curr} ) and the threshold value L _def1 are compared (S608). If the difference between the current propagation delay time and the previous propagation delay time (T _{delay_prev} -T _{delay_curr} ) is greater than the threshold value (T _def1 ), the node weights the weight for the hop number if the weight for the hop number is less than 1-Δ. By increasing (w1 = w1 + Δ), and if the weight for the number of hops is greater than 1- △, the weight for the number of hops is 1 (S610). If the weight of the surplus bandwidth is greater than Δ, the node decreases the weight of the surplus bandwidth by Δ (w2 = w2-Δ), and if the weight of the excess bandwidth is less than Δ, the weight of the surplus bandwidth is 0. (S610).

As a result of the step S608 determination, if the difference (T _{delay_prev} -T _{delay_curr} ) between the current propagation delay time and the previous propagation delay time is smaller than the threshold value T _def1 , the node may determine the weight w1 of the number of hops and the excess bandwidth. The weight w2 is maintained at the previous value (S612).

If it is determined that the traffic is increased (S614), the difference (T _{delay_curr} -T _{delay_prev} ) between the current propagation delay time and the previous propagation delay time is compared with the threshold value (T _def2 ) (S616). If the difference between the current propagation delay time and the previous propagation delay time (T _{delay_curr} -T _{delay_prev} ) is greater than the threshold value (T _def2 ), the node decreases the weight for the hop number by △ if the weight for the hop number is greater than △. (w1 = w1-Δ), if the weight for the hop number is less than △, the weight for the hop number is set to 0 (S618). If the weight of the surplus bandwidth is less than 1-Δ, the node increases the weight of the excess bandwidth by Δ (w2 = w2 + Δ), and if the weight of the excess bandwidth is greater than 1-Δ, the weight of the excess bandwidth Is set to 1 (S618).

As a result of the determination in step S616, when the difference between the current propagation delay time and the previous propagation delay time (T _{delay_curr} -T _{delay_prev} ) is smaller than the threshold value (T _def2 ), the node has a weight (w1) for the number of hops and The weight w2 is maintained at the previous value (S612).

After changing the weight, each node stores the currently measured propagation delay value T _{delay_curr} in a previous propagation delay value T _{delay_prev} and uses it in the next weight adjustment procedure. If the weight value is maintained as is through step S612, the currently measured propagation delay value T _{delay_curr} is not stored in the previous propagation delay value T _{delay_prev} . Steps S602 to S620 are periodically performed at each node, and each node may adaptively adjust a weight value.

7 and 8 are diagrams schematically illustrating a fourth example of adjusting a weight based on a traffic load according to an embodiment of the present invention. The fourth example infers the traffic load by using the gateway advertisement message delay from the Internet gateway router, and defines the weight value of the cost function as a table in advance according to the pre-classified delay time, and adjusts the weight value based on this.

FIG. 7 is a diagram schematically illustrating a detailed classification of a propagation delay time according to a network state according to an embodiment of the present invention. In FIG. 7, traffic loads between T _{delay_low} and T _{delay_high} are defined by being divided into N-1 steps. The T _{delay_low} and T _{delay_high} values may be set differently according to an implementation example based on measurement data or simulation results in an actual network.

Each node defines a weight value of a cost function in advance as a table according to a pre-classified propagation delay time as shown in FIG. 7, and adjusts the weight value based thereon. That is, when the propagation delay time is small, the weight of the hop number is greater than the weight of the excess bandwidth, and when the propagation delay time is the big, the weight of the excess bandwidth is larger than the weight of the hop number, thereby improving performance. When the propagation delay time is between T _{delay_low} and T _{delay_high} , the propagation delay time is divided into N-1 stages, and the weight of the cost function is appropriately adjusted.

8 is a diagram schematically showing a weight table according to a propagation delay time according to an embodiment of the present invention.

Each node uses the weight table as shown in FIG. 8 and the currently measured propagation delay time (T _{delay_curr} = T _node -T _gw ) to calculate the weight w1 of the hop number of the cost function and the weight w2 of the surplus bandwidth. You can decide. In the weight table of FIG. 8, each weight value may be set in advance through actual measurement or simulation, and may vary according to implementation examples. The value of function f can be set appropriately based on network traffic and performance. For example, set f _i (·) = i / N (stepped function) regardless of network load, or f _i (·) = (e ⁱ -1) / (e ^N -1) (exponential function) ), f _i (·) = log (i + 1) / log (N + 1) (log function).

Each mobile node periodically measures the propagation delay time, and calculates a cost function by selecting weights defined in the weight table according to the measured propagation delay time. Each node performs a handoff to the least expensive internet gateway router using the calculated cost function value. This process can be repeated periodically.

In the case of inferring the total network traffic using the minimum surplus bandwidth, each node can infer the total network traffic using the minimum surplus bandwidth and adjust the weight value of the cost function according to the inferred result.

The minimum surplus bandwidth of each node may be determined by various factors such as the amount of traffic in use and interference of the attention node. Approximately, there are various methods for estimating such surplus bandwidth. Examples of the prediction method may include a Listen technique or a Hello technique.

In the Listen technique, each node uses the Network Allocation Vector (NAV) to determine whether a channel is used or not. Find the percentage of available time. Also, in order to take into account the time when the channel cannot be used additionally, such as the time corresponding to the inter-frame space (IFS), the ratio of the available time to the channel is divided by the weight, and then the total bandwidth is multiplied by this value to minimize the minimum. Surplus bandwidth can be estimated.

In the Hello method, the Hello message of the Ad hoc On demand Distance Vector (AODV) protocol is applied, and each node collects bandwidth information used by itself and neighboring nodes within two hops. After subtracting the bandwidth used by the node within 2 hops from the total bandwidth, this value is divided by the weight to consider the influence of the control packet and the like to predict the minimum surplus bandwidth.

By using the above method, each node may extract the minimum surplus bandwidth currently available for itself. An implementer may select a prediction method according to an implementation example. Each node can infer the network traffic load from the mobile node to the Internet gateway router by using the minimum surplus bandwidth information updated by receiving the advertisement message from the Internet gateway router. For example, a low surplus bandwidth can indirectly infer that the network has a high traffic load.

A process of adaptive handoff using the predicted minimum surplus bandwidth as described above will now be described.

9 is a diagram schematically illustrating a fifth example of adjusting a weight based on a traffic load according to an embodiment of the present invention. 9 is a method of inferring a traffic load using a minimum surplus bandwidth, and changing the weight value of the cost function in accordance with the traffic change.

Each mobile node initially sets the hop number weight and the surplus bandwidth weight and changes the weight value according to the change of traffic. Each mobile node initially sets the hop number weight and the surplus bandwidth weight to w_hop_init and w_bw_int, respectively (S900). Each mobile node has an initial weight and uses it to calculate the cost function and perform handoff.

Each mobile node periodically measures surplus bandwidth information on the network (S902). Each node compares the currently measured surplus bandwidth (BW _curr ) with the previous surplus bandwidth (BW _prev ) (or the surplus bandwidth measured when the weight was most recently changed) (S904), and the current surplus bandwidth (BW _curr ). If it is large, it is determined that the traffic is reduced (S906), and if the previous surplus bandwidth (BW _prev ) is large, it is determined that the traffic is increased (S914).

If it is determined that the traffic is reduced (S906), the difference between the current surplus bandwidth and the previous surplus bandwidth (BW _curr -BW _prev ) and the threshold value (BW _def1 ) is compared (S908). If the difference between the current surplus bandwidth and the previous surplus bandwidth (BW _curr -BW _prev ) is greater than the threshold (BW _def1 ), the node increases the weight for the number of hops by △ if the weight for the number of hops is less than 1- △ (W1 = w1 + Δ), and if the weight for the hop number is greater than 1- △, the weight for the hop number is 1 (S910). If the weight of the surplus bandwidth is greater than Δ, the node reduces the weight of the surplus bandwidth by Δ (w2 = w2-Δ), and if the weight of the excess bandwidth is less than Δ, the weight of the surplus bandwidth is 0. (S910).

As a result of the determination in step S908, if the difference between the current surplus bandwidth and the previous surplus bandwidth (BW _curr -BW _prev ) is less than the threshold value (BW _def1 ), the node has a weight (w1) for the number of hops and a weight for the surplus bandwidth ( w2) is maintained at the previous value (S912).

If it is determined that the traffic is increased (S914), the difference between the current surplus bandwidth and the previous surplus bandwidth (BW _prev -BW _curr ) and the threshold value (BW _def2 ) is compared (S916). If the difference between the current surplus bandwidth and the previous surplus bandwidth (BW _prev -BW _curr ) is greater than the threshold (BW _def2 ), the node decreases the weight for the hop number by △ if the weight for the hop number is greater than △ (w1) = w1-Δ), if the weight for the number of hops is smaller than Δ, the weight for the number of hops is 0 (S918). If the weight of the surplus bandwidth is less than 1-Δ, the node increases the weight of the excess bandwidth by Δ (w2 = w2 + Δ), and if the weight of the excess bandwidth is greater than 1-Δ, the weight of the excess bandwidth Is set to 1 (S918).

As a result of the determination in step S916, if the difference between the current surplus bandwidth and the previous surplus bandwidth (BW _prev -BW _curr ) is less than the threshold value (BW _def2 ), the node has a weight (w1) for the number of hops and a weight for the surplus bandwidth ( w2) is maintained at the previous value (S912).

After changing the weight, each node stores the current surplus bandwidth (BW _curr ) in the previous surplus bandwidth (BW _prev ) (S920) and uses it in the next weight adjustment procedure. If the weight value is maintained as is through step S912, the current surplus bandwidth BW _curr is not stored in the previous surplus bandwidth BW _prev . Steps S902 to S920 are periodically performed at each node, and each node may adaptively adjust a weight value.

10 and 11 are diagrams schematically illustrating a sixth example of adjusting a weight based on a traffic load according to an embodiment of the present invention. In the sixth example, the traffic load is inferred using the minimum surplus bandwidth, and the weight value of the cost function is previously defined in a table according to the pre-classified surplus bandwidth, and the weight value is adjusted accordingly.

FIG. 10 is a diagram schematically showing a detailed classification of minimum surplus bandwidth according to a network state according to an embodiment of the present invention. In FIG. 10, traffic loads between BW _low and BW _high are defined by being divided into N-1 steps. The BW _low and BW _high values may be set differently according to the implementation example based on measurement data or simulation results in the actual network.

Each node defines a weight value of the cost function as a table in advance according to the surplus bandwidth pre-classified as shown in FIG. 10, and adjusts the weight value based thereon. That is, when the excess bandwidth is large, the weight for the number of hops is greater than the weight for the excess bandwidth, and when the excess bandwidth is small, the weight for the excess bandwidth is greater than the weight for the hop number to improve performance. When the surplus bandwidth is between BW _low and BW _high , the surplus bandwidth is divided into N-1 steps and the weight of the cost function is appropriately adjusted.

11 is a diagram schematically illustrating a weight table according to a propagation delay time according to an embodiment of the present invention.

Each node may determine a weight w1 on the number of hops of the cost function and a weight w2 on the surplus bandwidth by using the weight table shown in FIG. 11 and the currently measured surplus bandwidth BW _curr . In the weight table of FIG. 11, each weight value may be set in advance through actual measurement or simulation, and may vary according to an implementation example. The value of function f can be set appropriately based on network traffic and performance. For example, set f _i (·) = i / N (stepped function) regardless of network load, or f _i (·) = (e ⁱ -1) / (e ^N -1) (exponential function) ), f _i (·) = log (i + 1) / log (N + 1) (log function).

Each mobile node periodically measures surplus bandwidth, and calculates a cost function by selecting weights defined in a weight table according to the measured surplus bandwidth. Each node performs a handoff to the least expensive internet gateway router using the calculated cost function value. This process can be repeated periodically.

By using the adaptive handoff method as described above, it is possible to prevent limited network resource waste and increase efficiency of the wireless network.

The present invention is not limited to the above-described embodiments, and as can be seen in the appended claims, modifications can be made by those skilled in the art to which the invention pertains, and such modifications are within the scope of the present invention.

1 is a diagram schematically illustrating a network structure for connecting a MANET to the Internet

FIG. 2 is a diagram schematically showing traffic loads classified according to network conditions according to an embodiment of the present invention.

3 is a diagram schematically illustrating a first example of adjusting a weight based on a traffic load according to an embodiment of the present invention.

FIG. 4 is a diagram schematically illustrating a detailed classification of traffic load according to network conditions according to an embodiment of the present invention.

5 is a diagram schematically showing a weight table according to traffic load according to an embodiment of the present invention.

6 is a diagram schematically illustrating a third example of adjusting a weight based on a traffic load according to an embodiment of the present invention.

FIG. 7 is a diagram schematically illustrating a detailed classification of a propagation delay time according to a network state according to an embodiment of the present invention.

8 is a diagram schematically showing a weight table according to a propagation delay time according to an embodiment of the present invention.

9 is a diagram schematically showing a fifth example of adjusting a weight based on a traffic load according to an embodiment of the present invention.

FIG. 10 is a diagram schematically showing a detailed classification of minimum surplus bandwidth according to a network state according to an embodiment of the present invention.

11 is a diagram schematically illustrating a weight table according to a propagation delay time according to an embodiment of the present invention.

Claims (26)

In an ad hoc network, Setting an initial value of a hop number weight value and a surplus bandwidth weight value of the cost function at the node; Measuring a current traffic load on the network at the node; Comparing the current traffic load with a previous traffic load and comparing the difference between the current traffic load and the previous traffic load with a first set value if the current traffic load is large; If the difference between the current traffic load and the previous traffic load is greater than the first set value, decreasing the hop number weight value of the cost function and increasing the excess bandwidth weight value; Comparing the current traffic load with a previous traffic load and comparing the difference between the previous traffic load and the current traffic load with a second set value if the previous traffic load is large; And If the difference between the previous traffic load and the current traffic load is greater than the second set value, increasing the hop number weight value of the cost function and decreasing the excess bandwidth weight value. The method of claim 1, And if the difference between the current traffic load and the previous traffic load is less than the first set value, maintains the hop number weight value and the excess bandwidth weight value of the cost function. The method of claim 1, If the difference between the previous traffic load and the current traffic load is less than the second set value, maintaining the hop number weight value and the excess bandwidth weight value of the cost function. The method of claim 1, The cost function is

How to handoff in an ad hoc network. (only,

And

Are the weights for hop and weight for minimum surplus bandwidth,

And

Is the number of hops corresponding to the jth routing path from Internet gateway router i and the minimum surplus bandwidth, respectively) The method of claim 1, Wherein the node periodically measures the current traffic load on the network and calculates a cost function. In an ad hoc network, Measuring a current traffic load on the network at the node; Determining a hop number weight value and a surplus bandwidth weight value of a cost function using the weight table according to the current traffic load and the traffic load; And Handoff to the least expensive internet gateway router using the cost function with the weighted value determined. The method of claim 6, The cost function is

How to handoff in an ad hoc network. (only,

And

Are the weights for hop and weight for minimum surplus bandwidth,

And

Is the number of hops corresponding to the jth routing path from Internet gateway router i and the minimum surplus bandwidth, respectively) The method of claim 6, Wherein the node periodically measures the current traffic load on the network and calculates a cost function. In an ad hoc network, Setting an initial value of a hop number weight value and a surplus bandwidth weight value of the cost function at the node; Measuring a message propagation delay time on the network at the node; Comparing the current message propagation delay time with a previous message propagation delay time and comparing the difference between the current message propagation delay time and the previous message propagation delay time with a first setting value if the current message propagation delay time is large; If the difference between the current message propagation delay time and the previous message propagation delay time is greater than the first set value, decreasing the hop number weight value of the cost function and increasing the excess bandwidth weight value; Comparing the current message propagation delay time with the previous message propagation delay time and comparing the difference between the previous message propagation delay time and the current message propagation delay time with a second set value if the previous message propagation delay time is large; And If the difference between the previous message propagation time and the current message propagation delay time is greater than the second set value, increasing the hop number weight value of the cost function and decreasing the excess bandwidth weight value. Off way. The method of claim 9, The message propagation delay time is an ad hoc network that is a difference (T _node -T _gw ) between time information T _gw and a current time value T _node included in an advertisement message broadcast and received from an Internet gateway router. Handoff method. The method of claim 9, And if the difference between the current message propagation delay time and the previous message propagation delay time is less than the first set value, the hop number weight value and the excess bandwidth weight value of the cost function are maintained. The method of claim 9, And if the difference between the previous message propagation delay time and the current message propagation delay time is less than the second set value, the hop number weight value and the excess bandwidth width weight value of the cost function are maintained. The method of claim 9, The cost function is

How to handoff in an ad hoc network. (only,

And

Are the weights for hop and weight for minimum surplus bandwidth,

And

Is the number of hops corresponding to the jth routing path from Internet gateway router i and the minimum surplus bandwidth, respectively) The method of claim 9, Wherein the node periodically measures the current traffic load on the network and calculates a cost function. In an ad hoc network, Measuring a message propagation delay time on the network at the node; Determining a hop number weight value and a surplus bandwidth weight value of a cost function using the weight table according to the message propagation delay time and the message propagation delay time; And Handoff to the least expensive internet gateway router using the cost function with the weighted value determined. The method of claim 15, The message propagation delay time is an ad hoc network that is a difference (T _node -T _gw ) between time information T _gw and a current time value T _node included in an advertisement message broadcast and received from an Internet gateway router. Handoff method. The method of claim 15, The cost function is

How to handoff in an ad hoc network. (only,

And

Are the weights for hop and weight for minimum surplus bandwidth,

And

Is the number of hops corresponding to the jth routing path from Internet gateway router i and the minimum surplus bandwidth, respectively) The method of claim 15, Wherein the node periodically measures the current traffic load on the network and calculates a cost function. In an ad hoc network, Setting an initial value of a hop number weight value and a surplus bandwidth weight value of the cost function at the node; Measuring excess bandwidth on the network at the node; Comparing the current surplus bandwidth with a previous surplus bandwidth and comparing the current surplus bandwidth with the first set value if the current surplus bandwidth is large; If the difference between the current surplus bandwidth and the previous surplus bandwidth is greater than the first set value, increasing the hop number weight value of the cost function and decreasing the surplus bandwidth weight value; Comparing the current surplus bandwidth with the previous surplus bandwidth and comparing the current surplus bandwidth with the second set value if the previous surplus bandwidth is large; And If the difference between the previous surplus bandwidth and the current surplus bandwidth is greater than the second set value, decreasing the hop number weight value of the cost function and increasing the surplus bandwidth weight value. The method of claim 19, And if the difference between the current surplus bandwidth and the previous surplus bandwidth is less than the first set value, maintains the hop number weight value and the surplus bandwidth weight value of the cost function. The method of claim 19, And if the difference between the previous surplus bandwidth and the current surplus bandwidth is less than the second set value, maintains the hop number weight value and the surplus bandwidth weight value of the cost function. The method of claim 19, The cost function is

How to handoff in an ad hoc network. (only,

And

Are the weights for hop and weight for minimum surplus bandwidth,

And

Is the number of hops corresponding to the jth routing path from Internet gateway router i and the minimum surplus bandwidth, respectively) The method of claim 19, Wherein the node periodically measures the current surplus bandwidth on the network and calculates a cost function. In an ad hoc network, Measuring a current surplus bandwidth on the network at the node; Determining a hop number weight value and a surplus bandwidth weight value of a cost function using the weight table according to the current surplus bandwidth and the surplus bandwidth; And Handoff to the least expensive internet gateway router using the cost function with the weighted value determined. The method of claim 24, The cost function is

How to handoff in an ad hoc network. (only,

And

Are the weights for hop and weight for minimum surplus bandwidth,

And

Is the number of hops corresponding to the jth routing path from Internet gateway router i and the minimum surplus bandwidth, respectively) The method of claim 24, Wherein the node periodically measures the current traffic load on the network and calculates a cost function.

Images