KR20040107960A - Method for improving of mobile ad hoc networks - Google Patents

Abstract

PURPOSE: A method for improving an MANET(Mobile Ad-hoc NETwork) is provided to find out a path for an on-demand node and to minimize the cost expended in the process of path setup. CONSTITUTION: A user node requests a necessary path for communication(S701). In case that a desired destination node is in the range of the user node's management zone(S702), the user node transfers a packet to the destination node using an IARP(Intrazone Routing Protocol)(S703). Afterwards, the user node and the destination node update the available resource values stored in their respective IARP routing tables, based on the amount of available resources(S704). Meanwhile, if the destination node is out of the user node's management zone, the user node transfers a route request packet through a BRP(Broadcast Routing Protocol)(S705). The path node that delivers the route request packet subtracts the available resource value of the path node from the request resource value requested by the route request packet and judges whether the subtracted value is smaller than 0(S706). In case that the subtracted value is smaller than 0, the path node forwards the route request packet any more, marks a route response packet with 'QoS failed', and transmits it to the user node(S707).

Description

METHOD FOR IMPROVING OF MOBILE AD HOC NETWORKS

The present invention relates to a method for improving a mobile ad network, and provides a method of constructing a temporary network locally for communication when an existing infrastructure cannot be used for any reason or for efficient use of a communication band. The present invention relates to a method for improving a mobile ad network.

There have been tremendous advances in telecommunications and networking over the last few years, and communication services among mobile users in wireless environments are becoming increasingly popular. Also, with the development of laptop computers and wireless modems or wireless LAN devices, the wireless data communication market is expected to continue to develop. The development of wireless data communication suggests the possibility of development in a new field, and one of the areas that is attracting attention is the mobile ad hoc network (MANET: Mobile Ad Hoc Network).

There are two ways to enable communication between two users in a wireless environment. The first method is a method using an existing infrastructure (infrastructure), the two terminals find the nearest base station (base station) within their communication range and then connected through a wired network connected to the base station. Problems raised with this method usually appear in handoff. Guaranteeing to move between the two base stations without short delay and packet loss must be guaranteed and many techniques for this operation are under study. Another problem is that this ensures the correct operation of the user terminal only where the existing infrastructure is functioning properly. If the existing infrastructure does not work for some reason, the terminals cannot communicate at all.

On the contrary, the second possible communication method is to use an infrastructureless wireless network. Such a room is called an ad hoc network. In this case, mobile terminals communicate with each other by constructing a temporary network by themselves without the efforts of a structure or an operator. The ad hoc network is currently in the limelight as a supplement to the existing communication methods by providing a way to communicate by constructing a temporary network locally when the existing infrastructure cannot be used for any reason or for efficient use of the communication band. I am getting it.

In an ad hoc network, each mobile terminal needs to serve as a specialized router for delivering packets and maintaining connectivity to other terminals. However, many mobile terminals temporarily configure a network, and the network itself changes frequently, and because of the unstable environment due to frequent connection failures in a wireless environment, routing protocols used in wired environments are not suitable for ad hoc networks. have.

In general, a wireless ad hoc network means that mobile nodes form a temporary network without a pre-installed infrastructure. Each node has a wireless interface and communicates with each other using radio or infrared communication. An example of an ad hoc network is a laptop computer and a personal digital assistant (PDA) that communicate directly through the air interface. Nodes belonging to an ad hoc network may be mobile or stationary. 1 is a diagram illustrating a simple example of an ad hoc network, in which node A and node D are out of a range capable of communicating with each other. However, packets can be exchanged between Node B and Node C in the middle. At this time, Node B and Node C will act as routers in the four-node Ad hoc network.

An ad hoc network does not need any node with centralized administrative privileges. This method ensures that one of the nodes constituting the entire network moves and moves out of the communication range, or any node loses the communication capability, thereby bringing down the entire network. The nodes that make up mobile ad hoc have a limited transmission range, so they need to communicate through several hops in order to communicate with each other. Therefore, all nodes that want to join the ad hoc network must have the ability to forward packets to other nodes. Therefore, all nodes on an ad hoc network act as a host and a router. In addition, ad hoc networks must be capable of coping with network configuration changes and malfunctions at nodes. Changes or failures in these networks are fixed through network reconfiguration. For example, if a node goes out of network range and causes a link failure, the affected nodes will require a new path for the previously maintained path and solve the problem by finding another path for mutual networking. This operation results in a slight increase in transmission delay, but the network remains operational.

The characteristics of the mobile ad hoc network from other existing networks can be summarized as follows.

1. Dynamic Network Configuration: Nodes in an ad hoc network are mobile and can be dynamically and randomly connected to the network. As a result, the network configuration changes randomly over time and cannot be predicted. This characteristic also isolates the node from other networks. It may even consist of a divided network. In addition, unidirectional and bidirectional paths are mixed in the ad hoc network, and adjusting the transmission and reception parameters may affect the change of network configuration.

2. Resource Constraints: The biggest feature of mobile ad hoc networks is that they have limited resources.

This constraint may be that the link itself has or that of the host node itself.

• Bandwidth Limitation: Although there has been a significant increase in the bandwidth of wireless networks, there are still many shortages of radio resources compared to those available in fixed wired networks.

Mobile ad hoc networks often operate in heterogeneous networks with varying bandwidths and transmission delays. In addition, wireless links are subject to various errors such as fading, noise, interference, and path loss. Therefore, the actual bandwidth is much smaller than the maximum data rate of the radio channel. In practice, the received signal may be very poor or combined with several bits of error. Since mobile networks are an extended area of fixed wired networks, users want to continue to receive services used in wired environments even in wireless environments such as mobile ad hoc networks. The demand for bandwidth and link environment is increasing due to the increase of multimedia services that require wide bandwidth and the increase of various computing systems. However, these demands often cause congestion due to the small bandwidth and high error environment of the radio link, and sometimes they are always considered to be expected rather than exceptional.

Energy limitations. Mobile nodes rely solely on batteries as the source of power. Thus, effective development of systems and applications to reduce energy consumption adds to the complexity of ad hoc networks. Although battery technology has advanced a lot, there are still many shortcomings that are not enough to provide satisfactory performance in a wireless environment.

Restriction of Node's Capabilities: In mobile ad hoc networks, processors of each node are mounted in a small terminal. This limits the power of processing and also limits the data storage space. This environment means that traditional client server models or processing power at multiple application nodes must be reconsidered.

3. Limited physical security devices: Mobile wireless networks are less secure than conventional wired networks due to the possibility of device loss or transmission of data to channels that are not guaranteed to be secure. This is due to the dynamic environment of the mobile ad hoc network itself, which is not suitable for providing a secure security algorithm, and the circumstances of the node's limited processing capabilities. Therefore, careful efforts are needed to construct a safe room from various types of attacks such as eavesdropping, disguise, denial of service.

4. Autonomous network management: There are no centralized management objects in mobile ad hoc networks. All nodes must manage their own network configuration. The network runs on its own without any infrastructure.

The ad hoc network, meanwhile, began in the military research in the United States between the late 70's and early 80's. It was designed and developed to be used in emergency situations in order to prevent uncommunication in the military due to the destruction of command and communication facilities during military operations. It was developed mainly for disaster situations or military operations. However, due to the rapid development of various terminal devices and communication equipment, the range of use is expected to expand gradually. The scope of use that is currently being suggested is diverse, ranging from conference halls to document sharing and strengthening of the infrastructure used in the military. Representative examples are as follows.

1. Group Communications: Military applications are just one example of group communications. At any impromptu meeting, MANET can be used very simply. In such a situation, MANET enables ad hoc communication without setting up a special network configuration. It can also be used as an alternative when the existing infrastructure is not working for any reason.

2. Personal Communication: Personal Area Networking (PAN) means communication between devices that a user has around him. In the near future, a user may have a device capable of communicating everywhere, such as wallets, glasses, waistbands, hands, pockets. All these devices are located very close to the user to form the PAN. In communicating between these devices, MANET can make networking very simple and easy.

3. Application of Embedded Devices and Home Computing: In light of the current developments in communication technology, 'intelligent' systems will soon be introduced in TVs, refrigerators, air conditioners, coolers, and intrusion alarm systems. Currently, consumer electronics products using these technologies are being released. Ad hoc networks can be applied to communication between these devices.

However, in the successful introduction of the current MANET, first, the bandwidth that can be provided in the wireless environment is less than the wired environment, which will not change much in the future, so that the overall performance of the ad hoc network is greatly limited by this environment. There is this. Secondly, technologies and standards for wireless communication continue to increase, and each technology or standard focuses on a small part of the overall wireless environment to solve the problem. Not only does it provide a wide range of solutions to solve problems, but there is no common consensus on how these standards and technologies will interact with each other, and the factors that companies should consider in their business are playing their own role. Although similar problems can occur in MANET, only various solutions are proposed to solve the details of the big problem of ad hoc network. Third, due to the nature of the ad hoc network, communication must be performed through multiple nodes for data transmission, and there is a problem that a security loophole occurs because it is difficult to introduce a facility such as a key distribution method or an authentication server for security. Finally, user interest or disposition may have a profound effect on the future of ad hoc networks. That is, in the ad hoc network, it is assumed that all users play a common role for networking. Therefore, there is a problem in that it is difficult for a malicious user or an uncooperative user to join the ad hoc network.

2 is an exemplary diagram illustrating a difference between a traditional wired network and a mobile ad hoc network.

First, the traditional fixed Internet has a small mobility of the host or router, and works reliably. Mobile IP provides a technique for providing more flexible mobility to this fixed Internet, but still requires access to the wire curtain. As shown in FIG. 2, in the mobile ad hoc network, all nodes are mobile and each is a router or any part is connected to a wired network. The communication protocols operating in these mobile ad hoc networks must adapt adaptively, construct their own networks and meet the various constraints described in Section 2.2. Traditional routing protocols associated with wired networks cannot be used properly in mobile ac hoc networks because of severe resource waste and slow adaptability to network changes. Numerous protocols have been developed for mobile ad hoc networks, from packet radio networks in the Defense Advanced Research Project Agency (DARPA), the initial process of ad hoc networks. These networks have different characteristics and requirements from the existing networks in consideration of the characteristics of the ad hoc network, and the routing protocol used here must satisfy those characteristics and requirements. The proposed protocol for ad hoc networks has been proposed as a solution to some of the broad problems for the entire wireless network. To date, no routing protocol has been proposed that solves all problems for wireless networks. These problems are caused by various requirements for wireless networks and the ability of different terminals in different situations.

Routing protocols for fixed networks have been well studied and are currently used stably in various wired networks, including public Internet networks. The basic approach in MANET uses the same method as the wired network. Routing protocols required by MANET also include requirements for wired networks, such as detection and adaptation of network configuration changes, and stability and convergence. But there is a big difference between wired networks and MANET. In a wired network, events that are mostly static and dynamically change occur less often. MANET, on the other hand, is inherently always dynamic.

The routing function has three logical components. These are i) routing algorithms, ii) control messages and iii) routing data.

First, a routing algorithm is a logical structure that describes how routing can be achieved. The actual implementation of routing requires the exchange of some control messages, and how these messages are exchanged and what type of messages are exchanged is determined by the routing algorithm. The routing table refers to information data for maintaining and adding new information for routing. MANET adds the following constraints to the three logical elements for the routing functions described above.

Mobility affects routing algorithms.

Bandwidth limitations affect control message transmission.

Limitations of the node's capabilities affect the choice of routing data.

Since the existing wired routing protocol does not work properly in an ad hoc network, a routing protocol with the following characteristics is required in an ad hoc network.

Distributed operation: When designing a protocol, the operation of each node should be distributed. There is no centralized control node. The same applies to static nodes. Since each node is mobile, it must be free to enter and leave the ad hoc network.

Loop free: In order to improve the overall network performance, routing in routing protocols must satisfy the loop-free function. This avoids wasting bandwidth and CPU.

• On demand: The protocol must be operated on-demand to minimize the overhead of network management and prevent waste of network resources. This is because the protocol operates only on demand and is used without periodic brotocasting to prevent waste of resources in the wireless network.

One-way path support: Different parts of the network can form one-way paths due to differences in the transmission distances of nodes and various wireless environments. Support for these one-way paths as well as support for two-way paths can improve the performance of the entire routing protocol.

Security: The wireless environment has several security vulnerabilities and is particularly vulnerable to an impersonation attack. There is a need for some kind of preventive security mechanism to ensure more reliable transmission from the routing protocol. User authentication and data encryption can be suggested as an alternative to solve this problem, but the key distribution process and management in ad hoc networks can be problematic. As another method for security, a security scheme using IP-sec [4] using tunneling of all packets transmitted is also discussed.

Power Conservation: All nodes on an ad hoc network are devices that depend on limited battery capacity, such as laptop computers or PDAs. Therefore, it must support some kind of stand-by state for long lasting network. Several techniques have been proposed, and sleep-mode support is important in routing protocols.

· Multipathing support: To minimize the response of protocols to connection failures due to congestion and network conditions, protocols that support multipathing are encouraged. If one path does not work properly, ad hoc network efficiency can be increased when communication using alternate paths is possible immediately without finding a new path.

QoS support: Ad hoc networks need some kind of quality of service (QoS) support to successfully settle into future alternative communications and to use them efficiently. Support for this feature can greatly increase the coverage and frequency of ad hoc networks. For example, QoS support is essential to enable real-time multimedia transmission. To date, no ad hoc routing protocol is designed to support QoS. However, in view of the demands of multimedia applications and routing protocols, the ad hoc routing protocol will be proposed in the near future. In the ad hoc routing protocol, the most basic function for QoS is to find a path that satisfies QoS parameters in addition to finding the shortest distance and optimal path.

Many routing protocols have been proposed for MANET. Each routing protocol has been proposed to solve some problems associated with these logical structures. All algorithms use some form of flooding to construct the initial routing data. The protocol is classified according to the method of updating routing information. First, in a centralized algorithm, all routing settings are determined from one node of the network. On the other hand, distributed algorithms make routing calculations shared by all nodes on the network. On the other hand, the protocol can be classified in a static and adaptive manner by how the path change is determined according to the traffic input pattern. Static algorithms consist of source-to-purpose pairs whose paths are independent of the state of the traffic. In this case, the path changes only when the reference path or node does not work properly. This type of algorithm cannot produce high throughput on networks with widely varying traffic inputs. Most packet networks use some kind of adaptive way in which the source-destination path can vary depending on the state of traffic congestion. Also, as the most widely used method for ad hoc routing protocols, Proactive continues to try to maintain a path to all nodes in its network. Thus, this approach always manages routes that can be used immediately when the route to the destination is needed. Distance-Vector method is typically used. Reactive algorithms, on the other hand, route only as needed. Traditional flooding schemes fall under the category of reactive protocols.

One of the most important problems in designing routing protocols for ad hoc networks is that routing information is needed at least from neighboring nodes of the network structure that changes from time to time. In addition, as the number of nodes in the network increases, the number of nodes destined for the destination increases, thus requiring a large amount of information such as routing updates and routing tables. This traffic propagates through the wireless environment and consumes a considerable amount of resources.

In general, conventionally used routing protocols can be classified into two methods, reactive and proactive. The proactive protocol periodically updates the route information in the network, so that currently available route information continues to propagate to other nodes. Examples of such proactive protocols include distance-vector protocol families such as Wireless Routing Protocol (WRP) and Destination-Sequenced Distance-Vector (DSDV).

Reactive protocols, on the other hand, are routed only when there is a demand for a destination. Therefore, the route setting process proceeds only when a route to a destination is needed. Therefore, packets such as route requests are propagated throughout the network only when a route to a destination is needed. Examples of flooding reactive routing protocols include dynamic source routing (DSR), ad-hoc on demanded distance vector routing (AODV), and temporarily ordered routing algorithm (TORA).

The advantage of the proactive approach is that when a route to a destination is required, a route to that destination can be provided. On the other hand, since the reactive method does not have a path that can be used immediately, it propagates packets such as routing requests and routing responses throughout the network to find a path, and the delay may be quite large in some cases. In addition, the propagation of packets over the entire network causes a waste of radio resources. Such long delays and the transmission of excessive control packets often make pure reactive routing protocols unsuitable for real-time communications. However, due to the nature of the ad hoc network, where the network itself is constantly changing, the use of pure proactive routing methods results in the exchange of a large amount of control packets to maintain the path to nodes that are not needed, which is also a significant waste of radio resources. There is a problem that causes.

The present invention devised to solve the above problems is a routing protocol that can be proposed as a routing protocol suitable for an ad hoc network, and finds a path to a node required by on-demand and costs the route setting process. The purpose of the present invention is to provide a method for improving a mobile ad network.

1 is an exemplary diagram showing a simple example of an ad hoc network,

2 is a diagram illustrating a difference between a conventional wired network and a mobile ad hoc network;

3 is an exemplary diagram illustrating a structure of a zone routing protocol (ZRP) applied to a method for improving a mobile ad network of the present invention;

4 is an exemplary view showing an IARP zone configuration;

5 is an exemplary diagram illustrating query propagation of a BRP;

6 is a graph for a routing zone optimization process;

7 is an operation flowchart showing a resource reservation method in a method for improving a mobile ad network according to an embodiment of the present invention;

8A is an exemplary diagram illustrating a format of a control packet for IARP;

8B is an exemplary diagram showing a format of a control packet for IERP

9 is a flowchart illustrating a routing method in a method for improving a mobile ad network according to an embodiment of the present invention.

In order to achieve the above object, a method for improving a mobile ad network of the present invention may include: requesting a path required for communication at a user node intending to use the network; Determining whether the destination node with which the user node wants to communicate is within a management zone range of the user node; If the user node is within a management zone of the user node, forwarding the packet to the destination node using an IARP; And updating, by the user node and the destination node, the available resource values to respective internally implemented IARP routing tables with reference to the amount of available resources, wherein the control packet for the IARP includes QoS related parameters. The QoS related parameter may include a metric format and a metric value, wherein the metric value indicates the amount of available resources corresponding to each of the metric formats.

In addition, in order to achieve the above object, a method for improving a mobile ad network of the present invention includes a path setting method comprising: sending, by a user node, a packet classified according to QoS; And a path node processing the packet by a scheduling algorithm, respectively, wherein the scheduling algorithm is DWRR.

Hereinafter, the most preferred embodiments of the present invention will be described with reference to the accompanying drawings so that those skilled in the art can easily implement the technical idea of the present invention. .

3 is an exemplary view illustrating a structure of a zone routing protocol (ZRP) applied to a method for improving a mobile ad network of the present invention.

Zone Routing Protocol (ZRP) is a mixture of proactive and reactive methods. It uses a proactive method, but uses it within a limited range. A reactive approach is also used, but instead of routing through the propagation to all networks, a search across the network through a specific selected node is also used. In order for a routing protocol to be effective, changes in network configuration must have a local effect. The creation or change of a new link is important only in the local sense and should not have a significant effect on the whole network. Proactive routing protocols for the entire network will propagate such local changes as they consume a lot of resources throughout the network. ZRP limits this propagation only to neighbors managed as routing zones, thereby reducing the cost of propagating network configuration changes.

Referring to FIG. 3, NDP (Neighbor Discovery Protocol) is a MAC (Media Access Control) protocol that periodically broadcasts a 'Hello' beacon signal and receives a 'Hello' beacon signal from a neighbor to connect with a neighbor node. Used to recognize state. This connection state information is passed to the Intrazone Routing Protocol (IARP), a proactive routing protocol, and the IARP collects nodes with the minimum distance in hops to form a routing area no greater than a parameter called zone radius. do. Every node in the network maintains its own routing area and can overlap with the routing area of its neighbors. The IARP maintains link status information in its own area and is responsible for packet transmission in real time. When there is no destination node in the zone, it uses the Interzone Routing Protocol (IERP). It uses a message distribution service known as the Border Resolution Protocol (BRP), which is different from the existing broadcast-based route discovery for route discovery.

BRP transfers queries directly to peripheral nodes located at the boundary of the routing area, rather than an unconditional packet broadcast to existing neighbor nodes. This has the effect of reducing the amount of query traffic. The process of IERP is very similar to that of the existing route discovery protocol. The source node generates a query packet and relays the query packet through the out node determined by the bordercast algorithm, and the node receiving the packet checks whether the destination node is included in its area. If so, the route reply packet is sent back to the source to set the path, otherwise the query is bordercasted again.

Zone routing uses a proactive approach to managing nodes with local destinations. To be precise, nodes that exist at a distance no greater than the number of hops defined by the zone radius are defined as routing zones, and only those nodes manage the path using the proactive method. Every node has its own routing zone. It is noteworthy that the routing zone overlaps with the routing zones of neighboring nodes. A routing zone composed of a zone radius 2 around A is as shown in FIG. 4.

In Fig. 4, nodes B to F belong to the zone of A. Node G is located outside A's routing zone. The nodes that are two hops away from A are the neighboring nodes.

In order to configure a routing zone, each node needs to know which nodes are located around it. This information can be obtained through MAC (Media Access Protocol) if polling protocol is used. Otherwise, it can be obtained by implementing NDP (Neighbor Discovery Protocol). NDP is a simple protocol for broadcasting periodic Hello beacon signals. Receiving such a beacon, a node can recognize which node is located around it.

Information about these neighbors becomes basic information used for IARP. Each node maintains a route to its routing zone using a proactive method. The IARP can be used to modify the proactive link state protocol used in global environments such as OSPF. Any proactive routing protocol can be used as an IARP by allowing basic protocols such as OSPF to operate only in a limited range of routing zones and taking into account some modifications for ZRP. Basically, IARP's link state advertisement is limited to Time To Live (TTL) value and has the condition that IERP can be used with homeostasis.

ZRP uses IERP to find a path for communication with nodes other than nodes managed by IARP. IERP is a reactive protocol that is used on-demand only when there is no information destined for the destination. Any existing reactive protocol can be used for IERP.

In general, path discovery in a reactive routing protocol is divided into two processes: path request and path response. The route request process is initiated by a node that does not have information about the destination in its route table, which generates a route request packet and propagates it to its neighbors. If the neighbor node has a route to the destination, the route is notified to the originating node. Otherwise, the route request packet is continuously propagated. Subsequent route request packets destined for the same destination are considered redundant and are discarded. The route request packet is broadcast along the route of the network to find all possible routes to the destination. However, because of the unidirectional nature of the broadcast team, in some cases the path discovery process of nodes is redundant. In this case, the nodes may exchange route request packets requesting the same destination with their neighbors and perform duplicate route discovery. Therefore, the optimized route search process should be a way to exclude the search for the already discovered area spreading outward from the originating node. For this route search process, IERP optimizes the route search process using BRP (Broadcast Routing Protocol).

The advantage of the hybrid routing protocol used by ZRP is that query propagation can be propagated according to the known configuration of the zone. That is, the node receiving the route request packet can respond by representing its route to the nodes in its zone with reference to its route cache.

After the route to the destination is obtained, information about the configuration of its zone can be used to cope with link link failures or unoptimized routes. This result extends the lifetime of the path and reduces the length of the path, making the ad hoc network more stable with less delay in packet delivery. Out of radio propagation ranges to neighboring nodes can cause link failure and disrupt data flow using this path. In the case of pure reactive routing protocols, any path with a broken link will fail immediately and a new path discovery or recovery process must be performed to ensure end-to-end connectivity and a new path. Until a packet is found, data packets coming into the node are discarded or delayed, reducing the performance of the application.

In contrast, when using a routing zone, a node has information about all nodes in its zone as well as its neighbors, so that in the event of a link failure, a new path can be replaced immediately.

A node that needs a route to a destination first checks to see if the destination belongs in its routing zone. If the node is in a zone and there is a route to the destination, no further route discovery is required. On the contrary, if there is no known route to the destination, the originating node constructs a bordercast tree as its neighbor node and delivers a route query packet to the neighbor node destined for the neighbor node. The node receiving the route query packet checks whether it belongs to the bordercast tree of the query it forwards. If it turns out that it is not a member of its bordercast tree, it simply logs the receipt of the query and discards the message. If a node belongs to the node's bordercast tree, it checks to see if the destination of the query is in its routing zone, and if there is a route to the destination, sends a route reply to the originating node. Otherwise, the neighboring node constructs a bordercast tree and continues to forward packets. During this process, a node that has processed a routing query assumes its zone is covered to prevent the propagation of duplicate queries.

Referring to FIG. 5, assuming that a zone radius is two hops and node A has data to transmit to node L, node A first looks at whether node L is in its zone. In the example above, because node L is outside the range of A, node A constructs a bordercast spanning tree with its peripheral nodes D, E, F, and G to generate a route query. You pass this query to nodes B and C to pass. These nodes also look for L in their zones. Since node L is not found here, B and C also construct their own bordercast tree to pass queries. In the case of Node B, a query is sent to E and G to pass to the uncovered nodes F, H, and J among its peripheral nodes. C and M are excluded because they belong to A's routing zone. In the case of node G, since L is included in its zone, it no longer passes a query, but sends a route reply including A-B-G-J-L, which is found, to A. The entire bordercast path discovery process is shown in the following table.

According to the above table, the bordercasting method is more effective than the general flooding search method. If the network is point-to-point, the bordercast method generates five queries. On the other hand, when using the flooding method, 13 or 12 queries are generated.

The biggest feature of the zone routing protocol is that there can be a big difference in the configuration of the routing zone. If the routing zone is 1 hop, it works in the same way as the normal flooding reactive routing protocol. In general, the larger the range of the routing zone, the more the number of nodes that can be directly transmitted without additional path discovery increases the overall route discovery efficiency. However, the larger the routing zone, the greater the amount of data that needs to be managed proactively, which puts a heavy burden on nodes. The efficiency of path discovery and the management of routing zones are tradeoffs, and it can be said that configuring the optimal routing zone is the core of the ZRP structure.

To date, no mathematical formula is known to find the zone radius for optimal performance. Even if all the parameter values of the entire network are known, there is no direct mechanism for finding the optimal routing zone radius. A simple way to find the optimal zone radius as shown in FIG. 6 is to set the zone radius to minimize the amount of traffic added to the ZRP. Other methods include min-searching and traffic adaptive.

As shown in FIG. 6, in the ZRP, the IERP overhead is reduced, the IARP is increased, and the IERP traffic volume is larger and vice versa. If either IERP or IARP is larger than the other, the overall traffic of ZRP can be expected to increase rapidly. Therefore, it is expected that the amount of IARP and IERP of the optimal routing zone becomes 1: 1.

The method for improving a mobile ad network of the present invention, after all, relates to a method for improving a Quality of Service (QoS).

First, according to Recommendation E.800 of the Consultative Committee for International Telephony and Telegraphy (CCITT), the forerunner of the International Telecommunication Union (Telecommunication Standardization Sector), QoS means "a service that determines the degree of satisfaction of users who use the service. A collective result of performance. " In the rapidly growing wireless Internet market, multimedia technologies such as wired networks and VOD multicasting services, such as VOD multicasting services, should be applied to wireless networks in order to become a killer application in the wireless network. However, unlike wired network, there are three major limitations in guaranteeing QoS in MANET. ① Bandwidth limitation ② Dynamic network configuration of MANET ③ Limited processing power and storage capacity of mobile nodes.

The first is due to the limited radio channel nature of the MANET itself, and the second is that nodes, the most important feature of MANET, continue to move. That is, there is a difficulty in guaranteeing QoS due to connection establishment and resetting due to continuous change of location and frequent transmission failure of packets. And thirdly, unlike wired routers, all mobile nodes must be themselves transmitting and receiving nodes, while nodes in the middle must act as routers. However, unlike mobile wired routers, wireless mobile nodes have limitations in processing capacity and storage capacity. Therefore, additional processing for QoS should not overload the wireless node.

In general, the easiest way to guarantee QoS in network communication is to provide enough network resources of the network. In other words, in the case of a wireless network, sufficient bandwidth is secured, and an application throughput required by all users is processed by upgrading an air interface card. However, this method is difficult and difficult to realize in reality, and even if such a network environment is achieved, it will always be insufficient to meet the bandwidth requirements of the evolving application. Also, when all nodes use the network with the same transmission right, it is difficult to predict when congestion will occur due to user's transmission request. So the idea is "Network Traffic Engineering". The basic concept is to classify users according to service class and to access a network resource with a differential transmission right. Currently used network traffic engineering can be classified into two techniques.

Reservation-Based Engineering: This method reserves the network resources of the network in advance according to the QoS requirements of the application. The acceptance or rejection of resource reservation is determined by the bandwidth management policy. . This approach is currently used in Asynchronous Transfer Mode (ATM) or IntServ / RSVP techniques.

Reservation-less Engineering: Contrary to the above, no resource reservation scheme is used, and CAC (Connection) is a more advanced technique at each node that is not a flow reservation scheme for user-required applications. Admission Control, Policy Manager, Traffic Class, Queuing Mechanism are used to classify packets and process them according to service class. The current DiffServ model uses this technique.

In order to guarantee GoS in ad hoc network, various factors such as QoS MAC protocol, resource reservation method, and scheduling should be considered. Among the routing protocols we are interested in, the largest role for guaranteeing QoS can be found in finding a path that satisfies QoS parameters. For this purpose, the introduction of IntServ model is suggested as the closest alternative, but there are some limitations due to the characteristics of mobile ad hoc network. First, the huge computational load and storage space requirements of the IntServ model are a heavy burden for mobile nodes to handle. In addition, as the amount of flow increases, this burden becomes greater, which causes problems in scalability. Second, when using the IntServ model to reserve resources, additional network resources are wasted. In other words, in MANET, bandwidth is a limited resource that is not enough. Additional packets generated for resource reservation and verification of the IntServ model can waste bandwidth. Thirdly, the use of mechanisms such as QoS is required, but this also results in serious waste of network resources.

Accordingly, the present invention proposes a more simplified mechanism for QoS routing of ZPR. First of all, the resource reservation method using in-band signaling is used without the occurrence of additional signaling for resource reservation. This is done during the path establishment process through the existing IERP and IARP. Considering the characteristics, the function is very simplified compared to the function in the wired network. We also propose a simplified scheduling technique instead of introducing a mechanism such as CAC.

FIG. 7 is a flowchart illustrating a resource reservation method in a method for improving a mobile ad network according to an embodiment of the present invention.

First, a path required for communication is requested from a user node intending to use a network (S701).

Thereafter, it is determined whether the destination node to which the user node communicates is within the management zone range of the user node (S702).

If the destination node to which the user node communicates is within the management zone range of the user node, the user node forwards the packet to the destination node using IARP (S703). Herein, the control packet for the IARP includes a QoS related parameter, and the QoS related parameter includes a metric type and a metric value. In addition, the metric value indicates the amount of available resources corresponding to each of the metric types.

Thereafter, the user node and the destination node refer to the amount of available resources and update the available resource values in the respective IARP routing tables (S704).

On the other hand, if the destination node to which the user node wants to communicate is outside the management zone of the user node, the user node forwards the route request packet through the BRP (S705).

Thereafter, the route node that delivers the route request packet determines whether or not the value obtained by subtracting the available resource value of the route node from the requested resource value required by the route request packet is smaller than 0 (S706), If the value obtained by subtracting the available resource value of the path node from the required resource value required by the route request packet is less than 0, the route response packet is not propagated anymore and 'QoS failed' is indicated in the route response packet. After that, it transmits to the user node (S707).

On the other hand, if the value obtained by subtracting the available resource value of the route node from the requested resource value required by the route request packet is equal to or greater than 0, the route node forwards the route request packet toward the destination node. In operation S708, the available resource value stored therein is updated.

In other words, ZRP is composed of three parts: IARP, IERP, and BRP. Nodes within the zone managed by each node deliver packets using IARP, and nodes outside the zone find the route through IERP's route request and reply packet through BRP. Pass the packet. Each node manages the IARP and IERP routing tables, and the information in the IARP routing table is also used in the IERP routing table. The format of a control packet for IARP is shown in FIG. 8A.

In the proposed algorithm, in order to provide QoS service, each node is assigned a metric type. QoS related parameter values are stored in the metric and metric values. The types of metric types include bandwidth (0x31) and processing delay (0x32), and the metric value indicates the amount of resources currently available for each node of the type. Nodes send and receive these IARP packets to form the IARP routing table and record the quality of the path as well as the path information for the node. For a packet that is out of range, the IERP is used to find a path. In this case, the packet shown in FIG. 8B is used.

When sending this packet, if it is necessary to guarantee the QoS of the flow, the packet is transmitted by determining the type and resource amount to be supported by using this field. A node forwarding a route request subtracts its metric value from the value of its requested resource, and if the value is less than 0, it no longer propagates the packet, but a route reply Mark the packet as Qos Failed and send it to the source node (user node). If the subtraction value is greater than zero, the packet is continuously propagated to the destination node using IERP.

Another consideration here is that route replies by a node whose resource reservation using the actual IERP does not propagate from the actual originating node to the destination node have routing information going to their zone-scoped destination. Is sent. Thus, to ensure actual end-to-end QoS, even nodes with information about the destination propagate IERP to the destination mode without sending a route reply in advance. The destination node sends a Route Reply, and the node that receives this message reserves the resource for the required resource.

When applying a kind of selective resource reservation method QoS model using in-band signaling introduced to MANET, it can guarantee the quality of each flow through resource reservation for the path as desired. There is no guarantee of routing along the optimal link, and other general flows suffer from starvation problems by flows that guarantee QoS. Therefore, in order to solve this problem, each node uses a scheduling algorithm for each packet to minimize the above-mentioned problems. The Diffserv model is applied here and the QoS structure applied in the previous study is applied to each node.

FIG. 9 is a flowchart illustrating a routing method in a method for improving a mobile ad network according to an embodiment of the present invention.

First, in the ZRP model, the source node (user node) classifies and sends packets according to QoS (S801), and intermediate nodes (path nodes) process the packets by the scheduling algorithm (S802). Due to the environment that each node does not provide enough processing power and storage space for scheduling algorithms in ad hoc networks, it is difficult to apply the algorithms used by DiffServ routers to MANET as they are in large IP networks. FIFOs cannot provide QoS for high priority traffic and PQ can guarantee the best QoS (delay and jitter) for high priority traffic, but high priority traffic monopolizes the bandwidth, You will starve to death. WFQ is intended to overcome the shortcomings of PQ, but it is difficult to configure the hardware due to the large amount of computation in calculating packet start, arrival time and variable length packet size. WRR is proposed to overcome this problem, and this scheduling algorithm has many advantages. The first overcomes the limitations of FQ and PQ. That is, they support flows that require different bandwidths (that is, visit a high priority queue multiple times during one service round) and at least one packet is removed from the queue during each service round. Therefore, low-priority cues are also free from the problem of starvation. However, an ad hoc network does not handle cells of the same size as ATM, so if you process packets of variable length, you will not be able to allocate the correct bandwidth. In order to solve this problem, the WRR has been modified, and thus, an improved dynamic weighted round-robin (DWRR) scheduling algorithm can be applied to an ad hoc network.

The present invention described above is capable of various substitutions, modifications, and changes without departing from the technical spirit of the present invention for those skilled in the art to which the present invention pertains. It is not limited to the drawings shown.

The present invention is a routing protocol that can be proposed as a routing protocol suitable for an ad hoc network, and has the advantage of minimizing the cost spent in finding a path to a node required by demand (on-demand).

Claims (4)

In the improvement method of the mobile Ade network, Resource reservation method, Requesting a path necessary for communication at a user node intending to use the network; Determining whether the destination node with which the user node wants to communicate is within a management zone range of the user node; If the user node is within a management zone of the user node, forwarding the packet to the destination node using an IARP; And The user node and the destination node referencing the amount of available resources and updating the available resource values in respective internally implemented IARP routing tables. Including, The control packet for the IARP includes a QoS related parameter, the QoS related parameter including a metric format and a metric value, wherein the metric value indicates an amount of the available resource corresponding to each of the metric format. The improvement method of the mobile Ade network of the above. The method of claim 1, The user node forwarding a route request packet via a BRP if the destination node to which the user node wants to communicate is outside the management zone of the user node; Determining, at the path node delivering the route request packet, whether a value obtained by subtracting an available resource value of the path node from a request resource value required by the route request packet is less than zero; And If the value obtained by subtracting the available resource value of the path node from the required resource value required by the route request packet is less than zero, the path node no longer propagates the route request packet and fails through a route response packet. Transmitting to the user node The improvement method of the mobile Ade ｈｏｃ network comprising a. The method of claim 2, If the value obtained by subtracting the available resource value of the path node from the required resource value required by the route request packet is equal to or greater than 0, the path node forwards the route request packet toward the destination node, and internal Updating the available resource value stored in the The improvement method of the mobile Ade ｈｏｃ network comprising a. In the improvement method of the mobile Ade network, The route setting method is Sending, by a user node, classifying a packet by QoS; And Each path node processing the packet by a scheduling algorithm Including, The scheduling algorithm is DWRR The improvement method of the mobile Ade network of the above.