KR101913745B1 - Apparatus and method of configuring transmission route utilizing data plane application in software defined network - Google Patents

Abstract

A transmission path setting device for establishing a transmission path by utilizing a data plane application in a software defined network according to the present invention is disclosed. The transmission path setting device includes a controller configured to set a flow table related to a priority for each transmission path between the source and destination (Src-Dst). The transmission path setting device may further include an application module configured to perform an action having a largest compensation value (Reward) for each state based on the flow table. Thus, in accordance with the present invention, therefore, in a software defined network, the load concentrated on the controller can be reduced.

Description

[0001] Apparatus and method for establishing a transmission path utilizing a data plane application in a software defined network [0002]

The present invention relates to an apparatus and method for establishing a transmission path in a software defined network. More particularly, to an apparatus and method for establishing a transmission path utilizing a data plane application in a software defined network.

A software defined network is a networking technology that can implement and apply various network functions by separating control planes and data planes coexisting in network equipment such as existing routers. In a traditional network environment, expensive features such as firewalls, Deep Packet Inspection, and Traffic Monitors are required to be purchased and installed. Therefore, cost and maintenance There was a lot of burden on.

However, in the software defined network environment, as described above, since the function can be implemented / developed through the programming, the management technique applicable to the Internet of Things (Data Center Network) has been recently studied and developed. The overall structure of a software defined network can be divided into an application layer, a control plane, and a data plane. In this regard, Figures 1 and 2 illustrate the basic structure of an existing software defined network associated with the present invention.

However, in these existing software defined networks, many functions for network operation and management are placed on the open flow switch. In addition, the required functions of a specific field linked with the software definition network are located in the application layer and functions through connection with the controller. For this reason, in order to operate / manage a large-scale network including a large number of nodes with a controller alone, the controller may be overloaded and the network may not be operated efficiently.

Therefore, an object to be solved by the present invention is to reduce a load concentrated on a controller in a software defined network.

Another object of the present invention is to reduce an increase in transmission delay of user data in accordance with a load concentrated on a controller in a software defined network.

In order to solve the above problems, a transmission path setting apparatus for setting a transmission path using a data plane application in a software defined network according to the present invention is disclosed. The transmission path setting device sets a flow table related to priority for each transmission path between the source and destination (Src-Dst) when there are a plurality of output ports for one input port (In_port) ) Of the controller. The transmission path setting device may further include an application module configured to perform an action having a largest compensation value (Reward) for each state based on the flow table. Thus, in accordance with the present invention, therefore, in a software defined network, the load concentrated on the controller can be reduced.

According to one embodiment, the controller can receive a request message (Packet_IN Message) for the flow table for transferring the packet from the open flow switch when the packet flows from the host to the open flow switch. In addition, the controller may select a shortest-path multipath that can receive the request message and transmit the packet.

According to one embodiment, the open flow switch may utilize the forward direction ports in each flow table to determine the current state of the switch.

According to one embodiment, the application module can generate a reward table based on the transfer direction port information for the same input port in the flow table through a compensation value calculation module (R-calculator).

According to one embodiment, the flow table comprises: a flow entry configured to distinguish a plurality of flows from one another; A match field including a priority for each transmission route, an input port (in_port), a source address and a MAC address of the destination and an IP address for the flow entry; And an action field for the flow entry, the action field including an output port corresponding to the input port in_port.

According to one embodiment, the application module includes a status monitor (S-Monitor) module configured to monitor status associated with an increase in the number of transmit / receive bytes of ports other than the input port (In_port). The application module may further include a compensation value calculation module (R-calculator) configured to calculate a rate for each of the transmit / receive bytes and to monitor a status associated with the rate increase . In addition, the application module receives the state associated with the increase / decrease of the number of transmission / reception bytes and the increase / decrease of the respective rates, and performs an action having the largest compensation value for each state And may further include an action conductor module.

According to one embodiment, the status monitor module defines the corresponding state as 0 if the number of transmit / receive bytes at time t + 1 is equal to the number of transmit / receive bytes at time t, If the number of transmission / reception bytes in the transmission / reception byte is greater than the number of transmission / reception bytes at the time t, the corresponding state can be defined as 1 to monitor the status of the ports.

According to one embodiment, the action conductor module may optionally select a first action (a) from the state (s) when a packet entry is detected in the compensation value table Q (s, a) (S, a) to Q (s) by performing a first action (a) to observe the compensation value (r) and selecting a second action , a '). The action conductor module may repeat the compensation value observation and the compensation value table update until the reward table associated with the state s in the compensation value table is completed, A Patch decision can be performed by selecting an action having the largest compensation value through observation of repeated compensation values. At this time, if there is only one transmission direction port with respect to the input port, the update can be performed by increasing the priority of the corresponding flow entry.

According to an exemplary embodiment, the compensation value Reward may be determined according to a reciprocal of a transmission delay value measured at a corresponding port when the action conductor performs the action, and a value of a maximum bandwidth Can be calculated as the product of the weights that vary.

According to one embodiment, the action conductor module extracts the In port information of each flow entry in the flow table and extracts the available bandwidth of at least one output port corresponding to the input port And may receive the compensation value from the compensation value calculation module to determine a final action.

Also disclosed is an apparatus for establishing a transmission path utilizing a data plane application in a software defined network in accordance with another aspect of the present invention. The transmission path setting device includes a status monitor module configured to monitor a status associated with an increase in the number of transmission / reception bytes of ports other than the input port (In_port). In addition, the transmission path setting device may include an action conductor configured to perform an action having a largest compensation value (Reward) for each state by receiving a state associated with an increase in the number of transmission / reception bytes Action Conductor module.

According to one embodiment, the transmission path setting device includes a compensation value calculation module (R-Calculator) configured to calculate a rate for each of the transmission / reception bytes and monitor a status associated with the rate increase, As shown in FIG. At this time, the action conductor module receives the state associated with the increase of the number of transmission / reception bytes and the increase or decrease of the respective rates, and performs an action having the largest compensation value for each state Lt; / RTI >

According to one embodiment, the status monitor module defines the corresponding state as 0 if the number of transmit / receive bytes at time t + 1 is equal to the number of transmit / receive bytes at time t, If the number of transmission / reception bytes in the transmission / reception byte is greater than the number of transmission / reception bytes at the time t, the corresponding state can be defined as 1 to monitor the status of the ports.

According to one embodiment, the action conductor module performs an action having the largest compensation value Reward for each state according to the use of the ports and the received flow table Lt; / RTI > In addition, in the action conductor module, the state is defined according to whether or not the ports are used, and a compensation value from an action on the state may be varied depending on whether or not the ports are used.

Also disclosed is a method of establishing a transmission path utilizing a data plane application in a software defined network according to another aspect of the present invention. The transmission path setting method includes a status monitoring step of monitoring a status associated with an increase in the number of transmission / reception bytes of ports other than the input port (In_port). The transmission path setting method may further include a rate monitoring step of calculating a rate for each of the transmission / reception bytes and monitoring a status associated with the rate increase. In addition, the transmission path setting method receives the state related to whether the number of transmission / reception bytes is increased and whether the rate is increased, and performs an action having the largest compensation value for each state And an action execution step.

According to one embodiment, the status monitoring step defines the corresponding status as 0 if the number of transmit / receive bytes at time t + 1 is equal to the number of transmit / receive bytes at time t, If the number of transmission / reception bytes in the transmission / reception byte is greater than the number of transmission / reception bytes at the time t, the corresponding state can be defined as 1 to monitor the status of the ports.

According to an exemplary embodiment, the performing of the action may include, in a compensation value table Q (s, a) for the compensation value, a first action (a) for arbitrarily selecting a first action An action selection step; Observing the compensation value (r) by performing the first action (a); And a second action selection / compensation value table updating step of updating the compensation value table Q (s, a) to Q (s, a ') by selecting a second action a' from the state s can do.

According to an embodiment of the present invention, the action execution step may include: determining whether a reward value table associated with the state (s) among the reward value table is complete; And repeating the step of observing the compensation value and updating the second action selection / compensation value table until the compensation value table is completed, selecting an action having the largest compensation value through the repeated compensation value observation, And a path determining step of performing a path decision (Patch decision).

The transmission path setting method according to the present invention is advantageous in that the load concentrated on the controller can be reduced by disposing and processing functions such as a transmission path update and request in an application form in a data plane that was only responsible for packet transmission in a software defined network .

According to another aspect of the present invention, there is provided a transmission path setting method including arranging, in an application form, functions such as a transmission path update and a request in a data plane that has only received a packet in a software defined network, It is possible to reduce an increase in the transmission delay of the mobile station.

1 shows a basic structure of an existing software defined network related to the present invention.
2 is a conceptual diagram of an apparatus for setting a transmission path in a software defined network according to the present invention.
3 shows a basic operation procedure of reinforcement learning in connection with a method of setting a transmission path in a software defined network according to the present invention.
4 shows components of a transmission path setting apparatus for setting up a transmission path utilizing a data plane application in a software defined network according to the present invention.
FIG. 5 is a view showing states obtained by an application (or an application module) according to the present invention from an OpenVSwitch (open flow switch).
Figure 6 shows a table of compensation values at each input port in accordance with an embodiment of the invention.
FIG. 7 shows that a controller places a flow table on an OpenVSwitch based on a Packet_IN message, according to an embodiment of the present invention.
FIG. 8 shows a form in which a transmission path is set according to an embodiment of the present invention.
Figure 9 shows a flow diagram of a transmission path setup method for establishing a transmission path utilizing a data plane application in a software defined network according to the present invention.

BRIEF DESCRIPTION OF THE DRAWINGS The above and other features and advantages of the present invention will become more apparent from the following detailed description of the present invention when taken in conjunction with the accompanying drawings, It will be possible. The present invention is capable of various modifications and various forms, and specific embodiments are illustrated in the drawings and described in detail in the text. It is to be understood, however, that the invention is not intended to be limited to the particular forms disclosed, but on the contrary, is intended to cover all modifications, equivalents, and alternatives falling within the spirit and scope of the invention. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.

Like reference numerals are used for similar elements in describing each drawing.

The terms first, second, etc. may be used to describe various components, but the components should not be limited by the terms. The terms are used only for the purpose of distinguishing one component from another.

For example, without departing from the scope of the present invention, the first component may be referred to as a second component, and similarly, the second component may also be referred to as a first component. The term "and / or" includes any combination of a plurality of related listed items or any of a plurality of related listed items.

Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

Terms such as those defined in commonly used dictionaries are to be interpreted as having a meaning consistent with the contextual meaning of the related art and are to be interpreted as either ideal or overly formal in the sense of the present application Should not.

The suffix "module "," block ", and "part" for components used in the following description are given or mixed in consideration of ease of specification only and do not have their own distinct meanings or roles .

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, preferred embodiments of the present invention will be described with reference to the accompanying drawings. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS In the following description of the present invention, detailed description of known functions and configurations incorporated herein will be omitted when it may make the subject matter of the present invention rather unclear.

Hereinafter, a transmission path setting apparatus and method for setting a transmission path using a data plane application in a software defined network according to the present invention will be described. Here, the transmission path setting device may be a source / destination node as well as any node on the network including a plurality of nodes. Therefore, the transmission path setting device may be one of a network device and a router other than the user terminal.

2 is a conceptual diagram of an apparatus for setting a transmission path in a software defined network according to the present invention.

3 shows a basic operation procedure of reinforcement learning in connection with a method of setting a transmission path in a software defined network according to the present invention.

In connection with the method for setting the transmission path in the software defined network according to the present invention, transmission path setting by reinforcement learning can be performed. Therefore, the transmission path setting by the reinforcement learning will be described as follows.

Reinforcement learning is one of the learning methods belonging to the category of machine learning. In the environment to which the agent is applying, the agent observes the current state and maximizes the reward value or minimizes the penalty It is a way to learn by choosing the order of action or action. Based on the Markov Decision Process (MDP) model, it defines the states that the agent can obtain, the actions performed by the agents, and the rewards / penalties according to the actions, and proceeds through incremental learning procedures through interaction with the environment. FIG. 2 is a conceptual diagram of an apparatus for setting a transmission path in a software defined network according to the present invention. FIG. 3 is a flowchart illustrating a method of setting a transmission path in a software defined network according to the present invention. FIG. Referring to FIGS. 2 and 3, the transmission path setting apparatus 1000 for performing the basic operation procedure of the reinforcement learning includes an application module 100 and an open flow switch 200. Here, the application module 100 is referred to as an agent, and the open flow switch 200 may be referred to as a (runtime) environment module (Runtime Envirionment module).

3, the application module 100 receives information on states (State, St) from the open flow switch 200 and performs a specific action (At, At) based thereon. Based on this specific action (action, At), the open flow switch 200 delivers the compensation value (reward, Rt) to the application module 100. It is the ultimate goal of reinforcement learning to obtain the optimal compensation value by repeating the process of obtaining the compensation value based on such an action.

In relation to this reinforcement learning, SARSA (State-Action-Reward-State-Action) algorithm is one of the reinforcement learning algorithms. It executes an action in each state, and then receives the rewards and updates the Q table.

In this regard, Table 1 shows the basic concept of the SARSA algorithm according to the present invention.

That is, in the reinforcement learning process for obtaining a compensation value based on a specific action according to the present invention, Q is started by initializing a compensation value function (or a compensation value table) for all states s and actions a. Then, for each episode, the state set S is initialized and a specific action A is selected from the state set S using a policy from Q. [ Thereafter, a specific action A is performed to obtain the compensation value R and the updated state set S '. Next, an updated specific action A 'is selected from the updated state set S' using the policy from Q. As described above, the process of performing the selected action and obtaining the compensation value may be repeated until the state set reaches the terminal state. This termination state may correspond to a case where an enough necessary action is performed for each episode, or the obtained compensation value is converged or has a certain value or more.

Meanwhile, FIG. 4 shows components of a transmission path setting apparatus for setting a transmission path by utilizing a data plane application in a software defined network according to the present invention. That is, the data plane application structure for the routing function proposed in the present invention is shown. As shown in FIG. 2 and FIG. 4, the transmission path is set in the present invention by defining a flow table for the transmission path to be transmitted to the nodes on the data plane in the controller 10 as an action, and Reward To determine the action of the next state.

On the other hand, in the SDN environment, a monitoring function is implemented on the open flow switch, information about all the nodes is collected, and a transmission path is set according to the collected information. However, if there is a lot of work to be done by the controller 10 or a lot of parts to be managed, the monitoring function may be distributed to each node on the data plane because overloading may occur as the number of nodes in the data plane on the network increases.

As shown in FIG. 4, the transmission path setting apparatus 1000 includes an application module 100 and an open flow switch 200. Here, the application module 100 and the open flow switch 200 may be referred to as a DP application and an environment, respectively. In particular, a runtime environment may be an example of an environment.

The application module 100 also includes a state monitor module 110, a compensation value calculation module 120 and an action conductor module 130. Here, the status monitor module 110, the compensation value calculation module 120, and the action conductor module 130 may be referred to as S Monitor, R Calculator and Action conductor, respectively.

In addition, the open flow switch 200 includes a port status provision module 210 and a flow table module 220. Here, the port state providing module 210 and the flow table module 220 may be referred to as a physical port status (providing module) and a flow table (module), respectively. On the other hand, the flow table module 220 may also be referred to as an open flow switch including a flow table. In addition, an example of the port status providing module 210 corresponding to the physical port status (providing module) may be a network interface card (NIC).

Meanwhile, as shown in the left side of FIG. 4, the application module 100 and the open flow switch 200 may be referred to as Applicaiton and OVS (Open VSwitch), respectively. The State of the OpenVSwitch is determined by the port monitoring through the SR Monitor module. In the SR Monitor module, 0 (not in use) if Tx bytes, Rx bytes at time t + 1, Tx bytes at Rx bytes at time t, And a set of 0 or 1 values obtained by Output Ports is defined as a State, and when requested, is monitored and transmitted to the action conductor module 130. The compensation value Reward is calculated as a reciprocal of a transmission delay value measured at a corresponding port when the action conductor module 130 performs an action. In addition, the compensation value Reward may be a weighted value that depends on the reciprocal of a transmission delay value measured at a corresponding port when the action conductor performs the action, ). ≪ / RTI > This will be discussed in detail below.

With respect to the above description, the details will be described in terms of the application module 100 and the open flow switch 200 in FIG.

First, referring to FIG. 2 and FIG. 4, an overall transmission path setting process in which a packet is transmitted will be described as follows. On the other hand, this transmission path setting procedure is not limited to the following procedure, but can be changed within the scope of the present invention.

1) packet flows from the host to the open flow switch 200, the open flow switch 200 requests the SDN controller 10 for a multi-path flow table for transmitting the packet.

2) Request message (Packet_IN Message), the SDN controller 10 selects the shortest multipath that can transmit the packet. At this time, the shortest multipath is selected by simply using Hop Count in the flow table for multipath. Thereafter, the SDN controller 10 delivers the shortest multipath to each open flow switch 200.

3) Each open flow switch 200 receiving the flow table determines the current state of the switch by utilizing the ports of the output port in each flow table. The current state of the switch is achieved using a state monitored through the S-Monitor module 110. [

4) Also, a reward table is generated based on the transfer direction port information for the same input port in the flow table. The compensation value table is generated based on the compensation value calculated through the compensation value calculation module (R-calculator) module 120. Meanwhile, one example of the compensation value table may be the compensation value table shown in FIG. 6 to be described later, which will be described in detail below.

5) In this case, when there are a plurality of transmission direction ports for one input port, a port having the largest Reward value among the Reward values of each transmission direction port is selected and reflected in the flow table. Here, selecting a port having the largest Reward value corresponds to updating the corresponding transfer direction port to use the higher priority of the corresponding flow entry. This updating process is performed through the action conductor module 130. [

6) If there is only one transmission direction port for the input port, the action conductor module 130 also updates the corresponding flow entry by increasing its priority.

Meanwhile, the open flow switch 200 receiving the flow table, which is an aggregate of the first plurality of flow entries, performs the above process through the application. Therefore, based on the reward value in the current state, the efficiency of each port is verified and the final transmission path is selected.

The process of setting the transmission path according to the above-described process will be described in more detail based on the constituent elements.

The application module 100 is configured to transmit a flow table between the source and destination (Src-Dst) existing in the data plane in advance. At this time, when there are a plurality of output ports for one input port (In_port), the above-described operation can be performed. Here, the flow table is not provided by a request every time a packet transmission is required, and thus performs (and / or learns) an action using a flow table previously received prior to packet transmission and transmits the packet accordingly. Accordingly, the open flow switch 200 is configured to set a flow table regarding the priority for each transmission path between the source and destination (Src-Dst).

On the other hand, the application module 100 is configured to perform an action having the largest compensation value Reward for each state based on the flow table. In addition, the application module 100 is configured to transmit the packet by performing the action with the largest compensation value, and update the compensation value with the compensation value corresponding to the action. It should be noted that the packet can be transmitted not only from an edge node but also from an arbitrary node on the entire network. In addition, the packet may be transmitted to the destination through the flow table of the intermediate nodes on the route between the source node and the destination node.

The status monitor module 110 is configured to monitor a status associated with an increase in the number of transmission / reception bytes of ports other than the input port In_port. Specifically, the status monitor module 110 can monitor the Ethernet port of the hardware in which the OpenVSwitch is installed. In addition, in the case of such a monitoring, it is configured not to operate continuously but only when receiving the Packet_IN Message from the open flow switch 200. In other words, when the information is transmitted to each OpenVSwitch, it is operated by a request of the action conductor module 130, and a transfer port is determined and simultaneously enters a standby state. This is to prevent resource waste by monitoring at each node and to reduce the overall load of the data plane.

In this regard, the controller 10 transmits a request message (Packet_IN Message) for the flow table for transferring the packet from the open flow switch 200 to the open flow switch 200 when the packet flows from the host to the open flow switch 200 . In addition, the controller 10 may select the shortest-path multipath to which the packet can be transmitted in response to the request message. At this time, the shortest multipath is selected by simply using Hop Count in the flow table for multipath. Thereafter, the controller 10 transfers the shortest multipath to each open flow switch 200.

The open flow switch 200 can determine the current state of the switch by utilizing the transfer direction ports in each flow table. Also, the application module 100 may generate a reward table based on the transmission direction port information for the same input port in the flow table through a compensation value calculation module (R-calculator) 130. [ In one example, the compensation value calculation module 130 may be configured to calculate a respective rate for the transmit / receive bytes to monitor the status associated with the increase or decrease of the rate. Accordingly, the compensation value calculation module 130 may generate a compensation table based on the transmission direction port information and the rate information.

In this regard, the state of the OpenVSwitch corresponding to the flow table module 220 is determined by the port monitoring through the status monitor module 110. The status monitor module 110 monitors the statuses related to Tx bytes and Rx bytes at time t and t + 1 of the ports other than the input port through (1) and (2) in Equation 1, To the conductor module 130.

는 시간 t와 t-1 사이의 패킷 수신량에 해당하고,

는 시간 t와 t-1 사이의 패킷 송신량에 해당한다. 또한,

와

는 시간 t와 t-1에서 각 패킷 수신량에 해당한다. 또한,

와

는 시간 t와 t-1에서 각 패킷 송신량에 해당한다.here,

Corresponds to a packet reception amount between time t and t-1,

Corresponds to the amount of packets transmitted between time t and t-1. Also,

Wow

Corresponds to each packet reception amount at time t and t-1. Also,

Wow

Corresponds to each packet transmission amount at time t and t-1.

Concretely, 0 (not in use) when Tx bytes and Rx bytes at the time t and t + 1 are equal to 1 (in use) when it is increased can be represented. That is, the number of transmit / receive bytes corresponds to the number of accumulated bytes transmitted / received up to time t or t + 1 and equal to the number of transmit / receive bytes. It means that there is no data to be transmitted / received. Meanwhile, the status monitor module 110 defines the corresponding status as 1 if the number of transmission / reception bytes at the time t + 1 is greater than the number of transmission / reception bytes at the time t. That is, the number of transmit / receive bytes corresponds to the number of cumulative bytes transmitted / received up to time t or t + 1, and the increase in the number of transmit / receive bytes means that a new transmit / Means that there is data to be received. At this time, a set of 0 or 1 values obtained by the output ports is defined as a state, and the request is monitored and transmitted to the action conductor module 130.

As described above, if it is possible to define various states according to the presence or absence of newly transmitted / received data or other criteria, the method for defining the corresponding state is not limited thereto but can be freely modified.

In addition, the action conductor module 130 is configured to receive a state related to an increase in the number of transmission / reception bytes and perform an action having the largest compensation value Reward for each state . In addition, the action conductor module 130 receives the state associated with the increase in the number of transmission / reception bytes and the increase or decrease in the respective rates, and performs an action having the largest compensation value for each state . At this time, whether the number of the transmission / reception bytes is increased or not and whether the respective rates are increased are monitored from the status monitor module 110 and the compensation value calculation module 120, respectively, and the status related thereto is transmitted to the action conductor module 130 Lt; / RTI >

Here, the compensation value Reward may be calculated as an inverse number of a transmission delay value measured at a corresponding port when the action conductor module 120 performs the action.

In this regard, FIG. 5 is a view showing each state that an application (or an application module) according to the present invention can obtain from an OpenVSwitch (open flow switch). For example, as shown in Fig. 5, an application (or application module 100) can be obtained from OpenVSwitch (open flow switch 200) for eight states s (1) to s State of each state.

Meanwhile, in the action conductor module 130, a reward table for each state and action is generated through the above-described state, action, and compensation value (Reward). In the case of using the SARSA algorithm described above with reference to FIG. 3, the corresponding Q table is learned at the end of the episode and is used as a policy. However, in the method shown in FIGS. 4 and 5, since the state is defined by whether or not the port is used, the reward from the action may be different even if it is in the same state. For this reason, it is possible to apply algorithms that enable efficient action selection through continuous Q table update.

Specifically, this algorithm is an algorithm for selecting a transmission port in OpenVSwitch. Also, the difference from the existing reinforcement learning algorithm is that it is not an ε-greedy learning policy. In the proposed scheme, since the state is defined by the use of each port, the compensation value (Reward) from the action can be changed for each Packet_IN event even if it is in the same state. For this reason, it is possible to detect an incoming event and to select a transmission port through a state and an action at that time.

In this regard, Table 2 shows the pseudo code of the algorithm applied according to the scheme proposed in FIG. 4 and FIG.

That is, the action conductor module 130 arbitrarily selects the first action (a) from the state (s) when a packet entry is detected, in the compensation value table Q (s, a) (a), and observes the compensation value (r). In addition, the action conductor module 130 updates the compensation value table Q (s, a) to Q (s, a ') by selecting the second action a' from the state s. In addition, the action conductor module 130 repeats the compensation value observation and the compensation value table update until the reward table associated with the state s in the compensation value table is completed. The action having the largest compensation value can be selected through the observation of the repeated compensation value to perform a path decision. That is, the action conductor module 130 may be configured to perform an action having the largest compensation value Reward for each state according to the use of the ports and the received flow table. have. At this time, the state is defined according to whether or not the ports are used, and the compensation value from the action on the state can be changed depending on whether or not the ports are used. On the other hand, if there is only one transfer direction port for the input port, the update can be performed by increasing the priority of the flow entry.

A concrete scenario in which a packet is transmitted through the application proposed in the present invention is as follows. The transport path computed by the Packet_IN message is passed to each OpenVSwitch. At this time, the form of the flow entry is a pair of the In port and the output port, and a total of two flow entries are received for one transmission link. At this time, in order to find an efficient transmission port, a reward table is formed based on In port information on each match field. In the open flow switch 200, the source-destination transmission path is calculated by an algorithm based on the number of hops, and the priority of each transmission path is set differently and placed in the OpenVSwitch. Here, the Dijkstra algorithm can be used in connection with the hop-based algorithm described above for the source-to-destination transmission path. In other words, the Dextra algorithm receives a weighted graph G and a starting point s as inputs. Let V be the set of all vertices of graph G, and express the sides of the graph as a pair (u, v) of starting point u and destination point v. Let E be the set of all sides of G, and the weights of the sides are represented by the function w: E → [0, ∞]. In this case, the weight w (u, v) is the cost (time, distance, etc.) required to move from the vertex u to the vertex v. The cost of the path is the sum of the weights of all sides between the paths. The Dextra algorithm can be used to find the least costly path (shortest path) from s to t when there is a pair s and t of any vertex in V.

On the other hand, Table 3 below shows an example of the table for the flow entry between the source and destination disposed in one OpenVSwitch.

That is, the flow table may include a flow entry, a match field, and an action field. Here, the flow entry is configured to distinguish a plurality of flows from each other. The match field includes a priority, an input port, an origin and a destination MAC address (src_mac, dst_mac) and an IP address (dst_ip, src_ip) for each of the transmission paths with respect to the flow entry . In addition, the action field includes, for the flow entry, an output port corresponding to the input port (in_port).

Meanwhile, the action conductor module 130 extracts information on the In port of each flow entry in the flow table, and extracts, based on the available bandwidth of the at least one output port corresponding to the input port, And may receive the compensation value from the compensation value calculation module 120 and determine the final action.

Specifically, the action conductor module 130 extracts In port information of each flow table in the flow table disposed. This is the process that is performed to create the reward table based on the action rule information for the In port. Specifically, referring to Table 3, tables for in_port = 1, in_port = 2, in_port = 3, in_port = 4 are created. The table for in_port = 1 has Action Rule as output: 2, output: 3 and output: 4, so the available bandwidth of each output port for in_port = 1 in the current state And receives the compensation value Reward from the compensation value calculation module (R calculator) 120 to determine the final action. In the table for in_port = 2, in_port = 3, and in_port = 4, the flow entry is selected because it has only one action corresponding to output: 1. After determining the action according to the reward value in each reward table, the application module 110 can update the priority of the flow entry having the action with a high priority. In addition, in case of each entry, timeout time is applied, and if the entry is not utilized, it can be deleted after the timeout time. Here, the timeout time may be determined to be 5 seconds.

6 illustrates a table of compensation values at each input port in accordance with an embodiment of the present invention. 6, and as described above, there are at least one output port at each input port, for example, an action rule for in_port = 1 includes a plurality of output ports (output: 2, output: 3, and output : 4). That is, the flow entry is finally determined based on the compensation value Reward for each action. In this way, when each OpenVSwitch in the Data Plane updates the last flow entry, the packet is transmitted through the path based on the entries.

 In case of a flow entry determined through the reward table of in_port = 2, in_port = 3, in_port = 4, it is utilized or deleted according to whether a packet is transmitted from a neighboring node. This means that if a packet is delivered from a neighbor node connected to port 2, a flow entry having in_port = 3 and in_port = 4 as a match field is deleted by timeout setting and in_port = So that the packet is transmitted to the destination. This method determines the most efficient transmission path in the current state through the application of OpenVSwitch for Hop-based multi-path from the start point to the destination selected by the open flow switch. The most prominent reason for suggesting such a scheme is to exploit the more flexible transmission path setting method by actively utilizing the resources of the data plane in the SDN environment.

In the case of the compensation value Reward described above, the output port is extracted based on the information of the In port of each flow entry. In addition, when calculating a compensation value for each output port, the available bandwidth of the corresponding port is monitored and calculated and then transmitted to the action conductor module 130. (1) and (2) of Equation (1) described above, it is possible to confirm whether the Ethernet port is transmitting / receiving a packet, that is, the usage amount of the port. At this time, the SMonitor module monitors the transmission / reception bytes (Bytes) at the request of the action conductor module 130. At this time, if the transmitted and received bytes are increased, it is determined that the current port is being used and reflected in the state determination. Also, when each rate increases, the compensation value (Reward) is calculated by using it for the available bandwidth and transmission delay calculation.

Equations (3) to (5) are equations for calculating a compensation value (Reward) when an application performs an action on each reward table in the corresponding state . It utilizes the maximum bandwidth of each port to obtain the available bandwidth available and calculates the transmission delay based on the available bandwidth. The reason why (8) is multiplied by (3) and (4) in Equation 2 is to convert the Rate value obtained in byte units to Bit because the unit of bandwidth is Bit per Second.

와

는 각각 시간 t에서 포트의 가용 대역폭과 포트의 최대 대역폭을 나타낸다. 또한,

과

는 각각 시간 t에서 포트의 전송 지연 값과 유입된 패킷(Packet)의 크기(Size)를 나타낸다. 또한,

는 시간 t에서 해당 포트의 보상값을 나타낸다. here,

Wow

Represents the available bandwidth of the port and the maximum bandwidth of the port at time t, respectively. Also,

and

Represents the transmission delay value of the port and the size (size) of the incoming packet at time t, respectively. Also,

Represents the compensation value of the corresponding port at time t.

Equations (3) and (4) are formulas for calculating the reward value when the application performs an action on each reward table in the corresponding state. It utilizes the maximum bandwidth of each port to obtain the available bandwidth available and calculates the transmission delay based on the available bandwidth. The reason for multiplying 8 in Eqs. (3) and (4) is to convert the Rate value obtained in byte units to Bit because the unit of bandwidth is Bit per Second.

The Reward value is defined as the reciprocal of the transmission delay value, which can be obtained by performing packet transmission when the available bandwidth of each port is the same or different. Also, in the case of the weight, it is determined according to the value of the maximum bandwidth, which prevents the compensation value Reward from becoming large, and multiplies it to compare each port to an appropriate level. Based on the transmission delay value obtained through this process, the application performs each action and selects a port having the largest Reward value among the ports corresponding to the output, and transmits the packet.

In this regard, according to Equation (2), when the action conductor 120 performs an action, the compensation value Reward is represented as an inverse number of the transmission delay value measured at the corresponding port . On the other hand, the application module 100 may be configured to transmit a packet by performing an action having the largest compensation value, and to update the compensation value corresponding to the action. That is, the path that minimizes the transmission delay measured at the corresponding port is selected according to the packet transmission, so that the compensation value can be maximized.

FIG. 7 shows that a controller places a flow table on an OpenVSwitch based on a Packet_IN message according to an embodiment of the present invention. With respect to the multi-path selected in the controller, a total of five paths can be generated. That is, the path A passing SW1-SW2-SW3-SW4 and the path B passing SW1-SW4-SW5-SW6, the path C passing SW1-SW5-SW6, the path D passing SW1- SW2-SW5-SW6 to path E, and the flow table for this path is transmitted to OpenVSwitch in the path. Each entry is arranged after the controller selects multiple paths from the source to the destination, and several action rules are created for one In port in SW1, SW2, SW5, and SW6. It is an action that can be performed to transmit the packet that is coming from the current state of OpenVSwitch, and based on this, the Reward value is obtained by transmitting the packet based on the output port of each entry.

Each Flow Entry has a different priority. It has various Action fields in the same Match field. When the final action is determined, the priority is set to be higher than other entries, and the packet is used when it enters the In port in the Match field.

Meanwhile, FIG. 8 shows a form in which a transmission path is set according to an embodiment of the present invention. As shown in FIG. 8, among the actions that can be performed in SW1 and SW2, a flow entry having the largest Reward value is selected. Also, in the case of a flow entry which is not utilized by each OpenVSwitch, only the flow entry which is deleted by the timeout setting and used for the current packet transmission remains. When a packet is transmitted by another host after the route selection, the state is checked through the above-mentioned procedure, and a compensation value (Reward) according to an action to be performed is obtained, do. In the present invention, the most efficient packet transmission path setting is determined by the OpenVSwitch of the data plane without checking the current state and the efficiency of each port, and the transmission link is selected without further intervention of the open flow switch. Therefore, since the transmission link is selected without any additional intervention of the controller, the self routing organization technique can be specified.

In the foregoing, a transmission path setting apparatus for setting a transmission path using a data plane application in a software defined network according to the present invention has been described. Hereinafter, a transmission path setting method for setting a transmission path by utilizing a data plane application in a software defined network according to another aspect of the present invention will be described. Meanwhile, it goes without saying that the contents of the transmission path setting apparatus described above can also be utilized in the transmission path setting method.

In this regard, Figure 9 illustrates a flow diagram of a transmission path setup method for establishing a transmission path utilizing a data plane application in a software defined network according to the present invention. The transmission path setting method is performed by the transmission path setting device, in particular, by at least one of the application module 100 and the open flow switch 200. [

As shown in FIG. 9, the transmission path setting method includes a state monitoring step S100, a rate monitoring step S150, and an action performing step S200. Meanwhile, the action execution step S200 includes a first action selection step S210, a compensation value observation step S220, and a second action selection / compensation value table updating step S230. The action execution step S200 may further include a step S240 of determining whether the compensation value table is completed or not and a step S250 of determining a path.

In the status monitoring step S100, monitoring is performed to determine whether the number of transmission / reception bytes of the ports other than the input port In_port is increased.

In the rate monitoring step (S150), each rate is calculated for the transmit / receive bytes to monitor the status associated with the rate increase.

Meanwhile, in the state monitoring step S100, if the number of transmission / reception bytes at time t + 1 is equal to the number of transmission / reception bytes at time t, the corresponding state is defined as 0. Also, if the number of transmission / reception bytes at the time t + 1 is greater than the number of transmission / reception bytes at the time t, the corresponding status is defined as 1, and the status of the ports is monitored.

In the action execution step S200, a state associated with an increase in the number of transmission / reception bytes is received, and an action having the largest compensation value Reward is performed for each state.

Specifically, in the first action selection step S210, if the packet flow is detected in the compensation value table Q (s, a) for the compensation value, the first action a is arbitrarily selected from the state s.

In the compensation value observation step S220, the first action a is performed to observe the compensation value r. Accordingly, in the second action selection / compensation value table update step S230, the second action a 'is selected from the state s to set the compensation value table Q (s, a) to Q (s, a ').

Next, it is determined whether or not a compensation table associated with the state s in the compensation value table is complete in a compensation value table completion determination step S240. At this time, the compensation value observation step (S220) and the second action selection / compensation value table updating step (S230) are repeated until the compensation value table is completed.

 If it is determined that the reward table associated with the state (s) is completed, an action having the largest compensation value is selected through the repeated compensation value observation in the path determining step (S250) (Path decision).

According to the present invention, it is possible to reduce the load on the controller through the method proposed in the present invention on a network operated through a controller that performs other functions than routing. In addition, it is possible to reduce the load according to the flow table request even in a large-scale network including many nodes.

In connection with the expected effect of the present invention, it is possible to provide an environment that can reduce the load on the controller and smoothly perform other functions that the controller must perform by assisting the transmission path setting function through the method proposed in the present invention.

Regarding the commercialization prospect of the present invention, if the method proposed by the present invention is applied to a communication network utilizing a software defined network, it is expected that users can provide a more seamless network environment by reducing the load on the controller.

The transmission path setting method according to at least one embodiment of the present invention arranges and processes functions such as transmission path update and request in an application form in a data plane that has only performed packet transmission in a software defined network and processes the load, There is an advantage that it can be reduced.

In addition, a transmission path setting method according to at least one embodiment of the present invention is a method of setting a transmission path update and request in a data plane that has only performed packet transmission in a software defined network, It is possible to reduce an increase in the transmission delay of the user data depending on the load.

According to a software implementation, not only the procedures and functions described herein, but also each component may be implemented as a separate software module. Each of the software modules may perform one or more of the functions and operations described herein. Software code can be implemented in a software application written in a suitable programming language. The software code is stored in a memory and can be executed by a controller or a processor (or an application processor).

10: Controller
100: Application module
110: Status monitor module
120: compensation value calculation module
130: action conductor module
200: Open flow switch
210: Port status provision module
220: Flow Table Module
1000: transmission path setting device

Claims (16)

An apparatus for establishing a transmission path by utilizing a data plane application in a software defined network,
A controller configured to set a flow table regarding a priority for each transmission path between the source and destination (Src-Dst); And
When a plurality of output ports are provided for one input port (In_port), an action having the largest compensation value (Reward) is performed for each state based on the flow table And an application module. The method according to claim 1,
The controller comprising:
Receiving a request message (Packet_IN Message) for the flow table for transferring the packet from the open flow switch when a packet flows from the host to the open flow switch,
Selecting a shortest-path multipath to which the packet is to be transmitted by receiving the request message,
The open-
Determining a current state of the switch by utilizing the transmission direction ports in each flow table,
The application module comprising:
And generates a reward table based on the transfer direction port information for the same input port in the flow table through a compensation value calculation module (R-calculator). The method according to claim 1,
The application module comprising:
A status monitor (S-Monitor) module configured to monitor a status associated with an increase in the number of transmission / reception bytes of ports other than the input port (In_port); And
A compensation value calculation module (R-Calculator) configured to calculate a rate for each of the transmit / receive bytes, and to monitor a status associated with the rate increase; And
An action conductor module configured to perform an action having a largest compensation value for each state by receiving a state associated with an increase in the number of transmission / reception bytes and an increase or decrease in the respective rates, ). The method of claim 3,
The status monitor module comprising:
If the number of transmit / receive bytes at time t + 1 equals the number of transmit / receive bytes at time t, the corresponding state is defined as 0,
And when the number of transmission / reception bytes at the time t + 1 is greater than the number of transmission / reception bytes at the time t, the status is defined as 1 to monitor the status of the ports. The method of claim 3,
The action conductor module comprises:
In the compensation value table Q (s, a) for the compensation value, if the packet flow is detected, the first action (a) is arbitrarily selected from the state (s)
Performing the first action (a) to observe the compensation value (r)
The second action a 'is selected from the state s to update the compensation value table Q (s, a) to Q (s, a'),
Repeating the compensation value observation and the compensation value table update until the reward table associated with the state (s) in the compensation value table is completed,
If there is only one transmission direction port with respect to the input port, updating the flow entry with a higher priority,
Wherein the action having the largest compensation value is selected through the observation of the repeated compensation value, and a Patch decision is performed. The method of claim 3,
The compensation value (Reward)
Is calculated as a product of a reciprocal of a transmission delay value measured at a corresponding port when the action conductor performs an action and a weight depending on a value of a maximum bandwidth. and,
The action conductor module comprises:
Extracting an input port information of each flow entry in the flow table and outputting the compensation value from the compensation value calculation module based on an available bandwidth of at least one output port corresponding to the input port And determines a final action based on the received information. The method according to claim 1,
Wherein the flow table comprises:
A flow entry configured to distinguish a plurality of flows from each other;
A match field including a priority for each transmission route, an input port (in_port), a source address and a MAC address of the destination and an IP address for the flow entry; And
And an action field for the flow entry, the action field including an output port corresponding to the input port in_port. An apparatus for establishing a transmission path by utilizing a data plane application in a software defined network,
A status monitor module configured to monitor a status associated with an increase in the number of transmit / receive bytes of ports other than the input port (In_port); And
When a plurality of output ports are provided for one input port (In_port), a state associated with an increase or decrease in the number of transmission / reception bytes is received and a compensation value (Reward) And an Action Conductor module configured to perform an Action. 9. The method of claim 8,
Further comprising a compensation value calculation module (R-calculator) configured to calculate a respective rate for the transmit / receive bytes, and to monitor a status associated with the rate increase,
The action conductor module is configured to perform an action having the largest compensation value for each state by receiving a state associated with an increase in the number of transmission / reception bytes and an increase in the rate of each of the bytes, , A transmission path setting device. 9. The method of claim 8,
The status monitor module comprising:
If the number of transmit / receive bytes at time t + 1 equals the number of transmit / receive bytes at time t, the corresponding state is defined as 0,
And when the number of transmission / reception bytes at the time t + 1 is greater than the number of transmission / reception bytes at the time t, the status is defined as 1 to monitor the status of the ports. 11. The method of claim 10,
The action conductor module comprises:
An action having the largest compensation value Reward is performed for each state according to the use of the ports and the flow table received from the controller,
Wherein the state is defined according to whether or not the ports are used, and the compensation value from the action on the state depends on whether the ports are used or not. CLAIMS What is claimed is: 1. A method for establishing a transmission path utilizing a data plane application in a software defined network,
A status monitoring step of monitoring a status associated with an increase in the number of transmission / reception bytes of ports other than the input port (In_port);
A rate monitoring step of calculating a rate for each of the transmit / receive bytes to monitor a status associated with the rate increase; And
Receiving an increase in the number of transmission / reception bytes and a state associated with an increase in each of the rates, and performing an action having a largest compensation value for each state; How to set the path. 13. The method of claim 12,
The status monitoring step may include:
If the number of transmit / receive bytes at time t + 1 equals the number of transmit / receive bytes at time t, the corresponding state is defined as 0,
And if the number of transmission / reception bytes at the time t + 1 is greater than the number of transmission / reception bytes at the time t, the status is defined as 1 to monitor the status of the ports. 13. The method of claim 12,
Wherein the action-
A first action selecting step of arbitrarily selecting a first action (a) from the state (s) when a packet flow is detected, in a compensation value table Q (s, a) for the compensation value;
Observing the compensation value (r) by performing the first action (a); And
And a second action selection / compensation value table updating step of updating the compensation value table Q (s, a) to Q (s, a ') by selecting a second action a' from the state s , A transmission path setting method. 15. The method of claim 14,
Wherein the action-
Determining whether a reward value table associated with the state (s) in the reward value table is complete; And
The step of repeating the step of observing the compensation value and the step of updating the second action selection / compensation value table until the compensation value table is completed, selecting an action having the largest compensation value through observation of the repeated compensation value, Further comprising a path determination step of performing a decision (Patch decision). A computer-readable recording medium having recorded thereon a program for performing the transmission path setting method according to any one of claims 12 to 15.

Images