KR20080060882A - Method of adaptive routing based packet size and routing apparatus thereof - Google Patents

Abstract

An adaptive routing method and apparatus based on packet size are provided to set a route more properly and efficiently according to application or the packet size, thereby improving a transmission delay time or throughput to use a proper routing protocol according to the QOS(Quality Of Service) of each connection. A configuration of a node(200) having a routing protocol comprises a central processing unit(210), a routing processor(220), an application information extractor(230), a packet selection table(240), and a routing table(250). The central processing unit transmits received packets to the application information extractor if a routing request message is received. The application information extractor extracts application type information from a TCP/IP(Transmission Control Protocol/Internet Protocol) packet header, the received routing request message. The routing processor has the routing protocol, drivers the routing protocol to select a packet transmission route, searches the packet selection table to select packet size corresponding to received application information, and reflects the selected packet size to the routing protocol to set the packet transmission route. The routing table includes information about the packet transmission route.

Description

Adaptive packet size based routing method and routing device {Method of Adaptive Routing based Packet Size and Routing Apparatus}

1 is a diagram illustrating a configuration of a wireless mesh network.

2 is a diagram illustrating a configuration of a node equipped with a routing protocol according to the present invention.

3 is a schematic diagram illustrating a packet path routing process according to the present invention.

4 is a diagram illustrating a configuration of a TCP / IP packet.

5 is a diagram illustrating a result of measuring an average packet size for each application in a WLAN.

6 is a diagram illustrating a packet size selection table for each application.

7 is a diagram illustrating the configuration of a routing table for a path of a packet set according to the present invention.

8 is a diagram illustrating a process of driving a routing protocol according to the present invention.

9 is a diagram illustrating a process of resetting an existing path by reflecting a characteristic of a changed traffic.

             * Description of the symbols for the main parts of the drawings *

200: node 210: central processing unit

220: routing processor 230: application information extraction unit

240: packet selection table 250: routing table

The present invention relates to an adaptive packet size-based routing method and a routing device, and more particularly, dynamically changes a routing metric and a routing method according to an application on a wireless network, and further depending on the application. The present invention relates to a routing method and a routing device for performing routing based on a packet size by assigning a default packet size, and setting a new path accordingly when a characteristic of a transmitted traffic is changed.

Nodes having a function of switching packets among the components constituting the network apply various routing protocols to route the packets. Routing Protocol refers to a protocol for performing routing by implementing a specific routing algorithm. Examples of the routing protocol include RIP, OSPF, IGRP, and the like.

Routing Information Protocol (RIP), also known as route selection information protocol, routes using a distance vector algorithm that dynamically determines the shortest path based on the number of hops of the router.

Open Shortest Path First (OSPF) is designed to solve the problems of RIP. The OSPF (Open Shortest Path First) is designed to route the shortest path by combining distance information between nodes and link state information in real time.

The Interior Gateway Routing Protocol (IGRP) uses a distance vector algorithm and determines the route to the destination in consideration of factors such as packet delay, line bandwidth, reliability, and network load. In this case, the network administrator can prioritize these parameters.

Nodes refer to routing metrics when routing packets by applying the above routing protocol. Routing metrics are the number of hops to the destination (i.e. number of routers), bandwidth, communication cost, packet latency, network load, maximum transfer unit (MTU), path It may consist of factors such as cost and reliability. The routing metric value changes according to the factors constituting the routing metric and the routing metric calculation method, and accordingly, the path of the packet may be set differently.

Conventionally, nodes do not reflect the characteristics of an application in routing protocols when routing a packet. In other words, the routing of packets has been set in consideration of the network situation while excluding the possibility that the proper path may be different according to the characteristics of the application.

For example, in a wireless mesh network, a routing protocol that sets a path to have a minimum number of hops and a routing protocol that sets a path through a link having a maximum bandwidth are used. Let's assume that you apply and route.

In addition to sending packets, the control message transmission time (MAC overhead time) in the MAC is further increased while passing through one hop. The weight of the control message transmission time may vary depending on the size of the packet. In the case of transmitting a large packet, the MAC overhead time does not take a large portion, but in the case of transmitting a small packet, the MAC overhead time takes a large portion.

Therefore, if the packet size is large, even if the number of hops increases and the MAC overhead time increases, the weight of the packet will be relatively small. Therefore, it is advantageous to transmit the packet at a high data rate.

On the other hand, a small packet may have a larger portion of the MAC overhead time than the transmission time of the packet when the hop number increases and the MAC overhead time increases. Therefore, if the packet size is small, it is better to take the path with the minimum number of hops even over a low-rate link.

As a result, routing protocols that set the path through a minimum number of hops are suitable for applications that generate small packets, such as Telnet, and maximum for applications that generate large packets, such as File Transfer Protocol (FTP) services. Routing protocols that route through a link with bandwidth are appropriate.

Nevertheless, in the past, the characteristics of the application were not reflected in the routing protocol, and thus it was impossible to set an appropriate path for each application.

"High Throughput Route Selection in Multi-Rate Ad Hoc Wireless Networks" (2004), jointly proposed by Baruch Awerbuch, David Holmder and Herbert Rubens, reflects the weight of the link in the routing protocol. That is, according to this, weights according to link rate are used as routing metric, but packet size is assumed to be 1500byte to obtain this value.

Therefore, there is an error in the above study that the average packet size may be much smaller than the assumed 1500 bytes depending on the application. For example, the size of packets generated by applications such as telnet is much smaller.

Accordingly, an object of the present invention is to set an appropriate path by reflecting the characteristics of an application so that applications supported by a network can always have optimal performance. Based on the information, change how routing metrics are calculated. Accordingly, the present invention proposes a routing protocol for establishing a route.

Another object of the present invention is to create a path suitable for the application to improve transmission latency or throughput, ultimately using an appropriate routing protocol depending on the quality of service (QoS) of each connection. It provides a way to make it easy to add routing protocols when adding new services such as Voice over Internet Protocol (VoIP).

According to a packet routing method according to an application according to an aspect of the present invention for achieving the above object, the node examines the header of the received packet to determine the information of the application using the packet, accordingly the default packet size And calculating a routing metric by reflecting the default packet size and setting a path of the packet with reference to the node.

In the packet routing method according to the application, the application information is characterized in that the destination port number field of the header field of the path request message received by the node.

In the packet routing method according to the application, the step of selecting the default packet size may be selected by referring to a packet selection table that lists corresponding packet sizes for each application.

In the packet routing method according to the application, the routing metric, when there are a plurality of links constituting one path, divides the default packet size by the link bandwidth to determine the time required to transmit a packet on one link. After the calculation, the sum of the time required for each link is calculated.

In the packet routing method according to the application, if a difference between the average packet size obtained by measuring current traffic and the selected default packet size exceeds a reference value, the routing metric is recalculated to reflect the measured average packet size. And resetting the path with reference to the same.

In the packet routing method according to the above application, when a node receives a request for transmission of packets from a client, a default packet size is selected.

On the other hand, a node for routing a packet according to an application according to another aspect of the present invention for achieving the above object, the application information extraction unit and the application for extracting the application information by examining the header of the packet, and transmits it to the routing processor And a routing processor configured to search for the packet selection table listing the corresponding packet sizes for each packet, select a packet size corresponding to the application information, and set a path for transmitting a packet by driving a loaded routing protocol.

The node for routing the packet according to the application further includes a central processing unit for transmitting the received packet to the application information extracting unit when the path request message is received.

The node for routing a packet according to the above application further includes a routing table for constructing information on the path of the packet collected by the routing processor and providing the information on the routing.

In a node for routing a packet according to the above application, the routing processor is required to transmit a packet on one link by dividing the packet size by the link bandwidth when there are a plurality of links constituting one path. It calculates a time, calculates a routing metric by summing the time spent on each link, and then routes with reference to it.

On the other hand, according to the network according to another aspect of the present invention for achieving the above object, by identifying the application information from the client for transmitting the route request message to the node and the route request message received from the client to determine the corresponding packet size And a plurality of nodes configured to calculate a routing metric reflecting the selected packet size and to set a path for transmitting a packet with reference to the routing metric.

Hereinafter, an adaptive packet size based routing method and a routing device according to the present invention will be described in detail with reference to the accompanying drawings.

1 is a diagram illustrating a configuration of a wireless mesh network.

The wireless mesh network includes a mesh node and a mesh client. In the wireless mesh network, mesh nodes are wirelessly connected, and a mesh client is connected to one mesh node. In addition, the mesh client x 100 may access the Internet through the mesh node a.

The mesh node that is requested to set the path from the mesh client for the first time is called the originator node. In FIG. 1, a mesh node (f) and a mesh node (h), which are first requested to set paths from the mesh client x (100) and the mesh client y (200), respectively, become originator nodes.

Referring to FIG. 1, when the mesh client x 100 attempts to connect to the mesh client y 200, it can be seen that the paths of the packets that can be set are various. For example, connect to mesh client y (200) via a path consisting of nodes (f)-(g)-(h) from mesh client x (100) or node (f)-from mesh client x (100). It can connect to the mesh client y (200) via a path composed of (d)-(g)-(e)-(h). The nodes participating in this routing may vary depending on how the routing metric is calculated.

In order to perform appropriate routing for each application, the present invention proposes the following routing protocol. However, the following routing protocol is designed to apply the routing metric calculation scheme devised in the present invention, and the routing method according to the present invention is not limited to the routing protocol presented below. The following table compares the existing routing protocols with those of the present invention.

TABLE 1

The routing protocol proposed by the present invention is based on the radio-metric AODV (Ad hoc On demand Distance Vector) of IEEE 802.11s.

AODV calculates the routing metric based on the hop count. Thus, routing protocols that reference routing metrics based on AODV route packets to have a small number of hops. Therefore, when attempting to connect to the mesh client y (200) from the mesh client x (100) in Figure 1, a path through the mesh nodes (f)-(g)-(h) is set.

Radio-metric AODV, on the other hand, sets the path for better link performance. Therefore, a route with better bandwidth performance is chosen regardless of hop count, in which case the routing metric changes with AODV.

For example, in FIG. 1, the bandwidth of each link between node f and node g or a link composed of node g and node h is 2 Mbps, and the link between node f and node d is The bandwidth of each link consisting of the node d and the node g, the node g and the node e, and the node e and the node h is 11 Mbps. At this time, the radio-metric AODV selects a path consisting of nodes (f)-(d)-(g)-(e)-(h). In this case, the number of hops is increased compared to the case of the path consisting of mesh nodes (f)-(g)-(h), but the mesh nodes (f)-(d)-(g)-(e)-(h) This is because the bandwidth of the path consisting of the larger path has a larger bandwidth performance.

In addition, radio-metric AODV and AODV have different ways of handling duplicate route request messages (RREQ). RREQ message, a kind of routing control message, is a message for requesting routing of a packet to a node (or router).

AODV always discards duplicate RREQ messages (i.e., RREQ messages received later) if duplicate RREQ messages originating from the same originator node are broadcast. However, radio-metric AODV does not discard duplicate RREQ messages. This is to select a better packet path by comparing the path composed of the individual mesh nodes that broadcast the RREQ message.

As will be described later, since the routing protocol proposed in the present invention sets the path of a packet according to an application or a packet size, a path composed of respective mesh nodes that broadcast an RREQ message is compared. To do this, duplicate RREQ messages should not be discarded, so they are basically based on radio-metric AODV and not AODV. However, the method of routing packets is fundamentally different from that of radio-metric AODV. The routing protocol proposed in the present invention refers to the routing metric calculated by the following equation.

Equation (1) shown below is the most basic form of routing metric calculation.

……(1)

… … (One)

In Equation (1), P denotes a packet size to be transmitted, B denotes a bandwidth of a link, and O denotes an overhead time included in the transmission, including MAC overhead.

L _i is the routing metric of one link. L _i means a time required to transmit a packet on one link. Therefore, the routing metric (M) value is expressed as the sum of L _i when a set of links constituting one path is R.

If several paths can be set for one destination node, the routing processor selects the path with the smallest M value calculated by Equation (1).

However, it is considered that there is a substantial difference depending on the path of the packet that is set only when there is a difference that is greater than a threshold determined by the routing processor (this is called a routing metric threshold value (M _T )). This is because, if the difference is less than the routing metric threshold value (M _T ), the time difference for packet transmission is small by any path. Therefore, in the present invention, the routing metric threshold value (M _T ) is set to exclude the case of selecting a new route instead of a predetermined route for a very small advantage. The routing metric threshold value (M _T ) is obtained by experiment, and the value is 0.001.

Another type of routing metric calculation can be generated based on the above-mentioned routing metric calculation formula (1). In this case, the generated routing metric formula is composed of adding factors to consider. For example, when considering a factor called jitter to route on a path that can prevent data loss, the routing metric L _i of one link in the above equation is modified as shown in Equation (2) below. (In this case, however, the routing metric (M) calculation does not change.)

……(2)

… … (2)

If routing is performed considering the packet interval instead of the jitter factor, L _i is given by

……(3)

… … (3)

Now, the process of selecting a path according to packet size using Equation (1) will be described in detail.

In FIG. 1, when the packet is transmitted from the mesh client x (100) to the mesh client y (200), the packet node passes through a path composed of mesh nodes f)-(g) and mesh nodes f (f)-(d)-. There may be a case where a route consisting of (g) is used. In this case, the bandwidths of the links (f)-(g) are 2 Mbps, and the bandwidths of the links (f)-(d) and (d)-(g) are 11 Mbps.

For simplicity, it is assumed that there is no overhead in Equation (1). In this case, after node (f) broadcasts an RREQ message and it is received by nodes (d) and (g), node (d) again broadcasts an RREQ message and node (g) receives it. Let's say

The routing processor at node (g) must choose between the route generated via the RREQ message that arrived first and the route generated by the RREQ message that arrived later.

When the packet size is 60 bytes, when the routing metric M of the links f-g is calculated, 480bits ㆇ 2Mbps = 0.000240 (s). The routing metric (M) of links (f)-(d)-(g) is 480bits 11Mbps + 480bits 11Mbps = 0.000087 (s). The difference between these values is 0.000153 (s). Since this is less than the routing metric threshold value (M _T ) 0.001 set in the present invention, the path of the packet is not changed in this case. Node (g) sends a response message for RREQ message to node (f) which sent the first RREQ message, so that the packet is transmitted through the path consisting of nodes (f)-(g) along the existing path. .

If the packet size is 1400 bytes, the routing metric (M) of the link (f)-(g) becomes 11200bits ㆇ 2Mbps = 0.005600 (s), and the link (f)-(d)-(g) The routing metric (M) is 11200 bits s 11 Mbps + 11200 bits s 11 Mbps = 0.002036 (s). The difference between these values is 0.003564 (s). Therefore, in this case, the routing metric threshold value (M _T ) is greater than 0.001, so the routing processor changes the path of the packet. Node g sends a response message to the RREQ message to node d, which sent the later RREQ message, so that the path of the packet is changed to pass through nodes (f)-(d)-(g).

As a result, when the packet size is 60 bytes, the advantage of the new path is insignificant compared to the previously set path. Therefore, the path composed of nodes (f)-(g)-(h), which is the previously set path, is used as it is. If the packet size is 1400 bytes, the advantage of the new path will be greater, so choose a path consisting of nodes (f)-(d)-(g)-(e)-(h).

When routing according to the packet size is performed as described above, in the case of an application having a traffic characteristic having a small packet size, a path having the least hop is selected even though a link having a slightly slower transmission rate is used. In the case of an application having a traffic characteristic, a path having a higher number of hops but having the fastest transmission rate, that is, a link having the maximum bandwidth is selected. Therefore, when using the routing protocol proposed in the present invention, when the packet size (P) is small and operates similar to AODV, when the packet size is large, it operates similarly to the routing protocol that considers only the bandwidth differently.

2 is a diagram illustrating a configuration of a node equipped with a routing protocol according to the present invention.

The node includes a central processing unit 210, a routing processor 220, an application information extracting unit 230, a packet selection table 240, and a routing table 250.

When the route request message is received, the central processing unit 210 transmits the received packet to the application information extracting unit 230.

The application information extracting unit 230 extracts application information by inspecting the inside of the packet. That is, the application type information is extracted from the TCP / IP packet header which is the received route request message. In this case, the application information extractor 230 transmits the extracted application information to the routing processor 220.

The routing processor 220 includes a routing protocol, and selects a path to transmit a packet by driving the loaded routing protocol. In this case, the routing protocol mounted on the routing processor 220 is given in the same form as the above-described routing protocol.

       In addition, the routing processor 220 searches the packet selection table 240 to select a packet size corresponding to the received application information, and sets the path of the packet by reflecting it in the routing protocol. In this case, a routing entry for the corresponding route information is generated to update the routing table 250. The routing processor 220, which has updated the routing table 250, broadcasts a route request message to neighbor nodes.

       The packet selection table 240 lists the average packet size according to the destination port number.

       The routing table 250 contains information about the path of the packet. Factors constituting the routing table 250 may be a destination, an average packet size, a sequence number, a routing metric, and the like. The routing table 250 may be constructed by specifying an entry consisting of the above factors collected by the routing processor 220 and application information corresponding thereto.

3 is a diagram schematically illustrating a packet path routing process according to the present invention.

Prior to routing, the routing processor 220 sets default parameters of the routing metric referenced by the routing protocol (S301). In this case, default parameters that are applied according to the application are set differently. The application represents a type of service such as Telnet or FTP, and the default parameter is a factor to consider in routing, and may be a packet size, a bandwidth, and the like.

When the packet arrives, the application information extracting unit 230 checks the packet and obtains application information of the packet (S302). The obtained application information is transmitted to the routing processor 220.

The routing processor 220 calculates a routing metric by applying the application information (S303), and sets a path of the packet with reference to the routing metric (S304).

While the packet is being transmitted, the packet size may be changed according to the traffic characteristic. Therefore, the routing processor 220 periodically measures the average value of the packet size while transmitting the packet along the set path (S305).

In this case, the routing processor 220 determines whether the difference between the measured average packet size value and the default value is greater than or equal to the reference value (S306). If the difference is less than the reference value, the process is terminated without rerouting the packet.

However, if it is determined that the difference is more than the reference value, the routing processor 220 recalculates the routing metric based on the newly measured value (S307). Thereafter, the routing processor 220 re-routes the packet with reference to the recalculated routing metric (S308). This completes the routing process according to the present invention.

4 is a diagram illustrating a configuration of a TCP / IP packet.

The application information extracting unit 230 of the node 200 uses a service, i.e., application information, to be used by a subscriber from a protocol field located in an IP header and a destination port number field located in a TCP (or UDP) header. Know what is

In general, the protocol information field of the IP header contains information on which protocol is used, either TCP or UDP. Accordingly, the packet includes a TCP header or a UDP header.

That is, when the mesh client accesses the mesh node through TCP or UDP and requests the path setting, the application information extracting unit 230 may know that the packet uses the TCP or UDP protocol from the protocol information field of the IP header. The application information can be extracted by examining the destination port number of the TCP header or the UDP header included in the packet.

FIG. 5 is a diagram illustrating a result of measuring an average packet size for each application in a WLAN.

The figure lists the measured values by measuring the packet size that has an average for each application in a wireless LAN environment. For example, it can be seen from FIG. 5 that the size of an incoming packet is about 1500 bytes, the size of an outgoing packet is about 580 bytes, and the average packet size is about 580 bytes.

Since the measured values are obtained in a wireless LAN environment, different characteristics may result in different environments. Therefore, if packet size based routing is performed in the wireless mesh network environment, it is most reasonable to use the average packet size measured in the wireless mesh network environment.

In general, however, there are many applications that can predict the average packet size. For example, applications such as FTP generally send large packets, and applications such as chat send small packets.

6 is a diagram illustrating a packet size selection table for each application.

The packet size selection table is appropriately set by reflecting the average packet size measurement result for each application according to FIG. 5. The packet size selection table shown in FIG. 6 lists average packet sizes according to application types and destination port numbers.

In the present invention, the size of the packet is specified when there is a transmission request for packets, that is, at the beginning of transmission. Therefore, the default packet size is used as the packet size generally accepted according to the application.

For example, in case of Telnet, the default packet size is usually 50 bytes, and in case of FTP, the default packet size is 1400 bytes.

The routing processor 220 selects a default packet size P _r from the packet size selection table 240 using the application information and the destination port number. For example, referring to FIG. 6, if the destination port number is 23 as an application called telnet, the default packet size P _r is set to 60.

In the present invention, the application information and the destination port number are presented to find the packet size as described above, but this is only one example. Therefore, the information on the packet size can be obtained from other types of information depending on the system. For example, information about the packet size may be obtained from the contents of the packet rather than the header of the packet.

7 is a diagram illustrating a configuration of a routing table for a path of a packet set according to the present invention.

The routing table 250 contains information about the path of a packet. The routing processor 220 collects entries including information such as a destination (Dest), an average packet size (PktSize), a sequence number (Seq), a routing metric (metric), and the like. Then, the routing table is built by specifying the relevant application information.

Existing routing tables have only routing table entries configured, including destination nodes and next hops. For example, the existing routing table only contains information about the next hop to which packets from Node 9.

However, in the present invention, even if the destination is the same, different paths can be set according to the size of the packet, and thus have a routing table entry composed of information on the destination node, the packet size, and the next hop. This means that as the packet size varies, the routing table entries increase accordingly. In this case, the routing processor needs to search all routing table entries during routing, which can take a long time to establish the route.

After all, it is inefficient to configure the routing table by excessively varying the packet size. Therefore, the routing table should be configured by adjusting the packet size to an appropriate number.

8 is a diagram illustrating a process of driving a routing protocol according to the present invention.

The application information extracting unit 230 extracts application information and a destination port number by inspecting the header of the packet (S801).

The routing processor 220 selects a packet size P _r corresponding to the extracted information with reference to the packet size selection table 240 (S802). The selected packet size P _r is designated as the default packet size of the routing metric.

The routing processor 220 searches for whether an entry having <application information, a destination port number, and P _r > exists in the routing table 250 (S803). If an entry exists, the route of the packet is set using the route information according to the entry of the routing table 250 (S804).

However, if there is no routing table entry of the corresponding <application information, destination port number, P _r > in the existing routing table, the routing processor 220 must establish a new path. The routing processor 220 calculates a routing metric by substituting the packet size P _r into the parameter P of Equation (1) given above (S814).

The routing processor 220 sets a path of the packet with reference to the routing metric (S815). Thereafter, the routing processor 220 generates a routing table entry whose index is <application information, destination port number, P _r >, and updates the routing table 250 (S816).

Thereafter, the routing processor 220 determines whether a duplicate RREQ message has been received in order to determine whether the route needs to be reset (S805). If the RREQ message is not duplicated, the routing processor 220 terminates the packet routing process by the routing protocol.

However, if the RREQ message is received in duplicate, the routing metric value of each path consisting of the node transmitting the RREQ message is calculated (S806). The routing processor 220 compares the calculated routing metric values and determines whether the difference is greater than the routing metric threshold value (S807). If the routing metric threshold value is greater than 0.001, the path of the packet is reset to the path including the node transmitting the RREQ message later (S808). However, if the difference is less than 0.001, the routing process is terminated without rerouting the packet.

9 is a diagram illustrating a process of reconfiguring an existing path by reflecting a characteristic of changed traffic.

In the present invention, when the originator node receives a route setting request from the mesh client node, the originator node selects a default packet size according to information of an application to use the route, and sets a route according to the default packet size. However, since the size of a transmitted packet may be flexible, there may be a case where actual traffic is different from a predetermined packet size. Thus sending packets By measuring the average value over the packet size, If the average value is significantly different from the default value, the routing metric is recalculated to reflect the measured average packet size, and the existing path is reset. Such a processor consists of:

The routing processor 220 sets a threshold value P _T of the packet size difference, which is a criterion for determining that the characteristic of the traffic has changed (S901).

Thereafter, the routing processor 220 examines a packet size distribution of packets transmitted by the mesh client to measure an average packet size P _m (S902). In addition, the routing table 250 is searched to examine the packet size P _r currently used as the default packet size (S903).

The routing processor 220 determines whether the difference between the average packet size P _m and the packet size P _r currently being used is greater than or equal to the threshold value P _T of the packet size difference (S904). If the difference is less than the threshold value P _T of the packet size difference, the process of resetting the path is terminated.

However, if the difference is greater than the threshold value of the packet size difference, the routing metric is calculated using the measured average packet size P _m as a parameter and the packet is rerouted (S905). That is, the routing metric is calculated by changing the packet size P _r currently used as the default packet size to the newly measured average packet size P _m .

After rerouting the packet, the routing processor 220 determines whether a routing table entry having an index of <application information, destination port number, P _m > exists in the existing routing table 250 (S906). If the routing table entry exists, the routing table 250 does not need to be updated, thereby terminating the packet rerouting process.

However, if there is no routing table entry whose index is <application information, destination port number, P _m >, the routing table entry is generated to update the routing table 250 (S907). In this case, all rerouting of the packet ends.

The above description is for an example of setting the path according to the packet size. That is, the present invention proposes a fundamental method of mapping application information to a packet size and thus setting a path differently. In addition, the present invention also provides a method for resetting a path that is adaptive to the packet size, not limited to the default packet size. Therefore, it will be apparent to those skilled in the art from the above description that the path can be set according to packet size as well as various other factors, and further, the path can be reconfigured by adapting to the change of the factor.

In addition, the present invention has been described a method of applying a routing protocol according to the packet size using a wireless mesh network, the routing method proposed in the present invention can be applied to wired networks as well as various wireless networks.

According to the adaptive packet size-based routing method and routing apparatus according to the present invention, a more suitable and efficient route can be established by using an appropriate routing protocol according to an application or packet size, thereby improving transmission delay time or throughput. You can improve the quality of service (QoS) of each connection. Furthermore, by adding a new service, you can find the traffic characteristics (average packet size) and add a routing protocol for that service. There is an effect that can easily set the path.

Claims (11)

       In wired Internet Protocol (IP) networks,        Checking, by the node, a header of the received packet, identifying information of an application using the packet, and selecting a default packet size accordingly;        And calculating, by the node, a routing metric reflecting the default packet size, and setting a path of the packet with reference to the node. The method of claim 1, The application information, The packet routing method according to the application, characterized in that the node is identified from the destination port number field of the header field of the route request message received by the node.        The method of claim 2,        Selecting the default packet size,        The packet routing method according to the application, characterized in that the selection by referring to the packet selection table that lists the corresponding packet size for each application. The method of claim 1,        The routing metric,        When there are a plurality of links constituting one path, the default packet size is divided by the bandwidth of the link to calculate the time required to transmit a packet on one link, and then the sum of the time spent on each link is calculated. Packet routing method according to the application. The method of claim 1, If the difference between the average packet size obtained by measuring the current traffic and the selected default packet size exceeds the reference value, recalculating the routing metric reflecting the measured average packet size, and resetting the route with reference thereto Packet routing method according to the application further comprising.        The method of claim 1, When a node receives a request for transmission of packets from a client, it selects a default packet size.        An application information extracting unit which extracts application information by inspecting a header of the packet and transmits the application information to the routing processor;        A packet selection table listing packet sizes corresponding to applications; And And a routing processor configured to search the packet selection table to select a packet size corresponding to application information, and to drive a loaded routing protocol to set a path for transmitting a packet.        The method of claim 7, wherein If the route request message is received, the node for routing the packet according to the application further comprising a central processing unit for transmitting the received packet to the application information extraction unit.        The method of claim 7, wherein Node for routing the packet according to the application further comprises a routing table for building information about the path of the packet collected by the routing processor, the routing table for providing information about the route. The method of claim 7, wherein The routing processor, If there are multiple links that constitute one path, the packet size is divided by the bandwidth of the link to calculate the time it takes to transmit the packet on one link, and the routing metric is calculated by adding the time spent on each link. The node for routing the packet according to the application, characterized in that for routing with reference to this.        A client sending a route request message to the node; Wow        A plurality of nodes that grasp application information from a route request message received from a client, select a packet size corresponding thereto, calculate a routing metric reflecting the selected packet size, and set a route for transmitting a packet by referring to the routing metric. Network comprising.

Images