KR101554477B1 - Method, appratus, system and computer-readable recording medium for adjusting flow against overload - Google Patents

Abstract

An embodiment of the present invention relates to a packet processing method for a network switch. The packet processing method includes the following steps: (a) the network switch detects whether the network switch is overloaded or not; and (b) the network switch bypasses a packet to a neighbor network switch adjacent to the network switch when the network switch cannot confirm the route of the predetermined packet transmitted to the network switch alone while the network switch is detected to be overloaded. The present invention bypasses the packet required to be processed to be influenced by the overload of the network switch to the neighbor network switch adjacent to the network switch, thereby preventing problems due to the overload of the network switch.

Description

METHOD, APPARATUS, SYSTEM AND COMPUTER READABLE RECORDING MEDIUM FOR ADJUSTING FLOW AGAINST OVERLOAD BACKGROUND OF THE INVENTION Field of the Invention [0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method, an apparatus, a system and a computer-readable recording medium for processing a packet in response to an overload. More particularly, To a method, apparatus, system, and computer-readable recording medium for bypassing a packet whose path can not be determined by itself to a neighboring network switch adjacent to the network switch.

In general, the interface of the network device is determined by the manufacturer of the network device. The network switches of various manufacturers can provide different application programming interfaces (APIs). Since various network devices providing the same or similar functions provide different interfaces according to the manufacturer or version, the operator of the network may have difficulties in adding, developing or configuring functions according to the environment of the network. Thus, a unified interface to the network device, for example a standard API or an open interface, is required.

Software-Defined Networking (SDN) is a technology that is standardized by the Open Network Foundation (ONF). In SDN, network routing, network control, and other network operations are handled by software programming.

SDN means a network structure in which the control functions of the network are separated from the physical network. That is to say, the SDN is divided into a data plane and a control plane of the network, and a standardized interface between both planes is provided. For example, an operator of a network can use the interface provided by the SDN via programming on the control plane, and control the communication function on the data plane according to the context of the network via the interface provided by the SDN .

Among the functions provided by the SDN, there are software defined forwarding. In an existing network, packet forwarding is handled by the hardware of the switch or router. Software defined forwarding means that such packet forwarding is controlled by open interfaces and software. OpenFlow is widely used as a standard interface for software-defined forwarding. Openflow is an open source API defined to support programming of switch and router functions, providing flow control, security, and virtualization for SDN.

As elements constituting the SDN, there are the SDN controller and the SDN switches. The SDN controller uses the same protocol as the OpenFlow protocol to control the SDN switch. For example, the SDN controller may set the forwarding method of the packet according to the network topology, and may transmit the set forwarding method to the SDN switch.

The SDN switch can process packets under the control of the SDN controller. The SDN switch can use a flow table to process packets. For example, each flow entry in the flow table may include Match fields, Counters, and Action, and may define how to process packets that meet certain conditions. For example, if an outport action is specified in a flow entry, the packet may be forwarded to the outport.

In some cases, a particular packet may not be defined by the flow table. That is to say, a particular packet may not match any entry in the flow table. The SDN switch sends an upcall to the SDN controller for flow setup for this packet. The SDN controller generates a flow entry matched with the packet in consideration of the topology of the network and sends the information of the flow entry generated as a response to the up-call to the SDN switch. The SDN switch processes the packet using the newly provided flow entry for the packet.

As described above, the SDN switch processes packets using software. Therefore, the SDN switch may have a problem in that it uses a lot of resources in the processing of a packet as compared with a switch that processes packets by hardware. In particular, when the SDN switch is overloaded, the operation according to the above-described SDN principle can greatly affect the performance of the SDN switch. For example, if a packet sent to an overloaded SDN switch does not match the flow entries in the flow table, then many resources are required for the SDN switch to process the response to the upcall and upcall, Is lowered.

The present inventors have developed a technique that enables a load of a network switch to be reduced by detecting an overload of the network switch and bypassing a packet requiring processing that is affected by the overload to a neighboring network switch adjacent to the network switch It came to the following.

It is an object of the present invention to solve all the problems described above.

The invention also pertains to detecting overload of a network switch and allowing a reduction in the overload of the network switch by bypassing a packet requiring processing that is affected by the overload to a neighboring network switch adjacent to the network switch, .

The present invention also provides a method and apparatus for enabling the fulfillment of a Quality of Service (QoS) for a packet by bypassing a packet from an overloaded network switch to a network switch that is not overloaded, .

In order to accomplish the above object, a representative structure of the present invention is as follows.

According to an aspect of the present invention, there is provided a method of processing a packet of a network switch, comprising the steps of: (a) detecting whether the network switch is overloaded; And (b) if the network switch is in a state where it is detected that the network switch is overloaded, if the network switch is unable to determine the path of a predetermined packet transmitted to the network switch alone, And a step of bypassing the packet with a switch.

According to another aspect of the present invention, there is provided a network switch comprising: an overload detection unit for detecting an overload of the network switch; And if the network switch is in a state in which it is detected that the network switch is overloaded, the network switch can not determine the path of a predetermined packet transmitted to the network switch alone, There is provided a network switch including a packet processing unit for passing packets.

According to still another aspect of the present invention, there is provided a method for controlling a network switch, comprising a plurality of network switches, wherein a first network switch of the plurality of network switches detects whether the first network switch is overloaded, The first network switch bypasses the packet to the second network switch when the first network switch is in a state where the first network switch can not determine the path of the predetermined packet transmitted to the first network switch alone, Wherein the first network switch is a network switch neighboring the first network switch among the plurality of network switches.

In addition, there is further provided another method, apparatus, system for implementing the invention and a computer readable recording medium for recording a computer program for executing the method.

According to the present invention, it is possible to prevent a problem caused by an overload of a network switch by bypassing a packet requiring processing that is affected by an overload of the network switch to a neighboring network switch adjacent to the network switch.

Further, according to the present invention, the effect that the QoS for the packet can be satisfied by bypassing the packet from the overloaded network switch to the overloaded network switch is achieved.

1 is a diagram schematically showing a configuration of a network system according to the present invention.
2 is a diagram illustrating a configuration of a network switch according to an embodiment of the present invention.
FIG. 3 is a flowchart for explaining overload detection of a network switch and packet processing after detection of an overload according to an embodiment of the present invention.

The following detailed description of the invention refers to the accompanying drawings, which illustrate, by way of illustration, specific embodiments in which the invention may be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice the invention. It should be understood that the various embodiments of the present invention are different, but need not be mutually exclusive. For example, certain features, structures, and characteristics described herein may be implemented in other embodiments without departing from the spirit and scope of the invention in connection with an embodiment. It is also to be understood that the position or arrangement of the individual components within each disclosed embodiment may be varied without departing from the spirit and scope of the invention. The following detailed description is, therefore, not to be taken in a limiting sense, and the scope of the present invention is to be limited only by the appended claims, along with the full scope of equivalents to which such claims are entitled, if properly explained. In the drawings, like reference numerals refer to the same or similar functions throughout the several views.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings, so that those skilled in the art can easily carry out the present invention.

Configuration of the entire system

1 is a diagram schematically showing a configuration of a network system according to the present invention.

As shown in FIG. 1, the entire system of the present invention may include a network controller 110 and a plurality of network switches. The network controller 110 may be an SDN controller, and each network switch of the plurality of network switches may be an SDN switch.

In FIG. 1, a first network switch 121, a second network switch 122, a third network switch 123 and a fourth network switch 124 are shown as a plurality of network switches. In the network, the control path is shown in dashed lines and the data path is shown in solid lines.

According to one embodiment of the present invention, each network switch of the plurality of network switches may comprise one or more ports. In Figure 1, each port is shown as a small rectangle for convenience. Each port of the network switch can receive packets from another network switch or transmit packets to another network switch. In addition, each port of the network switch may receive packets from an external network or an external device, or may transmit packets to an external network or an external device. In a packet processed by a network switch, the inbound port may refer to a port of the at least one port of the network switch that has received the packet, and the outbound port may be a port of at least one of the network switches It may be the port that sent the packet.

Network switch configuration

2 is a diagram illustrating a configuration of a network switch according to an embodiment of the present invention.

According to an embodiment of the present invention, the network switch 200 may include an overload detection unit 210, a packet processing unit 220, a storage unit 230, and at least one port 240. According to an embodiment of the present invention, at least some of the overload detection unit 210, the packet processing unit 220, and the storage unit 230 may be program modules communicating with an external system. Such program modules may be included in the network switch 200 in the form of an operating system, application program modules, and other program modules, and may be physically stored on various known memory devices. Also, at least some of these program modules may be stored in a remote storage device capable of communicating with the network switch 200. Such program modules may include routines, subroutines, programs, objects, components, and / or components that perform particular tasks or perform particular abstract data types, Data structures, and the like, but are not limited thereto.

According to one embodiment of the present invention, each network switch of the plurality of network switches described above with reference to FIG. 1 may correspond to the network switch 200 of FIG. Each network switch of the plurality of network switches may include an overload detection unit, a packet processing unit, and at least one port. The overload detection unit, the packet processing unit, the storage unit, and the at least one port of each network switch are connected to the overload detection unit 210, the packet processing unit 220, the storage unit 230 and the at least one port 240, Respectively.

According to an embodiment of the present invention, the overload detection unit 210 may sense an overload of the network switch 200, and may detect an overload state after the overload is detected.

According to an embodiment of the present invention, the packet processing unit 220 may perform processing on a packet. The packet processing unit 220 can receive a packet through at least one port and can transmit a packet through at least one port. In addition, the packet processing unit 220 may transmit the upcall to the network controller 110 for processing the packet, and may receive a response to the upcall from the network controller 110.

According to an embodiment of the present invention, the storage unit 230 may store data required for operation of the network switch 200. [ For example, the storage unit 230 may store a flow table of the network switch 200, bypass information, and the like.

In accordance with one embodiment of the present invention, at least one port 240 of the network switch 200 may be coupled to another device within the network system. Other devices within the network system may include a network controller 110 and other network switches. Also, another portion of the at least one port 240 may be coupled to a device external to the network system.

Overload detection and processing of packets after overload detection

FIG. 3 is a flowchart for explaining overload detection and packet processing after detection of an overload in a network switch according to an embodiment of the present invention.

Hereinafter, in the network system described above with reference to FIG. 1, a first one of a plurality of network switches 121 is described as an overloaded network switch, and a second one of the network switches 122 Is described as a network switch.

(i) First, according to an embodiment of the present invention, the overload detection unit 210 of the first network switch 121 may sense the overload of the first network switch 121 (S310). The overload detection unit 210 of the first network switch 121 may determine that the first network switch 121 is overloaded when the load of the first network switch 121 is equal to or greater than a predetermined threshold .

According to one embodiment of the present invention, the overload of the first network switch 121 may be controlled by a switch-controller tcp (TCP) or a switch-to-controller (SSL) session which is the control path of the first network switch 121 - Controller Session). That is to say, the overload of the first network switch 121 may mean a session between the first network switch 121 and the network controller 110 or an overload in the processing related to the session.

According to an embodiment of the present invention, in order to detect an overload of a network switch, a usage amount of a central processing unit (CPU) (hereinafter referred to as "CPU") of the network switch, Congestion, and the number or rate of flow entries in the flow table of the network switch. The overload detection unit 210 of the first network switch 121 detects the usage amount of the CPU of the first network switch 121 and the usage amount of the CPU of the first network switch 121, Overload of the first network switch 121 based on at least one of the number or rate of flow entries in the flow table. For example, the overload detection unit 210 of the first network switch 121 may determine that the first network switch 121 is overloaded when the usage amount of the CPU of the first network switch 121 is greater than a predetermined value . Alternatively, the overload detection unit 210 of the first network switch 121 may determine that the first network switch 121 is overloaded when congestion between the first network switch 121 and the network controller 110 is greater than a predetermined level . Alternatively, the overload detection unit 210 of the first network switch 121 may determine that the first network switch 121 is overloaded when the number or ratio of the flow entries in the flow table of the first network switch 121 is equal to or greater than a predetermined value It can be judged.

According to one embodiment of the present invention, congestion between a network switch and a network controller may be determined by the number or ratio of pending upcall flow entries in the network switch. For example, congestion between a network switch and a network controller may be proportional to the number or proportion of pending upcall flow entries in the network switch. For reference, in the specification of the present invention, a flow entry, which is an entry of a table used for packet forwarding, and an upflow flow entry, which is an entry of a table waiting for uplink, are used separately.

(ii) Next, in accordance with an embodiment of the present invention, when an overload of the first network switch 121 is detected, the packet processing unit 220 of the first network switch 121 performs overloading of the first network switch 121 To the network controller 110 (S320).

(iii) Next, in accordance with an embodiment of the present invention, when an overload of the first network switch 121 is notified, the network controller 110 determines whether a packet among the plurality of network switches adjacent to the first network switch 121 The neighboring network switch to be bypassed can be selected (S325). Herein, a network switch adjacent to the first network switch 121 may mean at least one network switch capable of reaching the first network switch 121 in, for example, one hop (Hop) in the network topology. As shown in FIG. 1, the network switches adjacent to the first network switch 121 are the second network switch 122, the third network switch 123 and the fourth network switch 124, 2 network switch 122 is determined as a neighboring network switch to which the packet will be bypassed.

According to one embodiment of the present invention, the network controller 110 is configured to determine the traffic statistics of each network switch of a plurality of network switches adjacent to the first network switch 121, the amount of CPU usage and the number or percentage of flow entries in the flow table Based on at least one of the plurality of network switches adjacent to the first network switch 121, the neighboring network switch to which the packet will be bypassed. For example, the network controller 110 may select a network switch having the smallest traffic statistics among a plurality of network switches adjacent to the first network switch 121 as a neighboring network switch to which the packet will be bypassed. Alternatively, the network controller 110 may select a network switch having the lowest CPU usage among a plurality of network switches adjacent to the first network switch 121 as a neighboring network switch to which the packet will be bypassed. Alternatively, the network controller 110 may select a network switch having the smallest number or percentage of flow entries in the flow table among the plurality of network switches adjacent to the first network switch 121 as a neighboring network switch to which the packet will be bypassed .

According to one embodiment of the present invention, the network controller 110 may generate bypass information by selecting a neighboring network switch to which the packet will be bypassed. The bypass information may be information indicating a neighboring network switch to bypass the packet. The bypass information may indicate a selected neighboring network switch (that is, a second network switch 122, or a port connected to a selected one of the at least one port of the first network switch 121).

(iv) Next, according to an embodiment of the present invention, the network controller 110 may transmit the bypass information to the first network switch 121 (S330). The packet processing unit 220 of the first network switch 121 may receive the bypass information from the network controller 110. [

(v) Next, according to an embodiment of the present invention, bypass information may be set for the data path of the first network switch 121 when the bypass information is transmitted (S335). For example, the overload detection unit 210 of the first network switch 121 may set the bypass information for the data path of the first network switch 121. Alternatively, the packet processing unit 220 of the first network switch 121 may also perform a set of bypass information. In addition, the set of bypass information may indicate that the first network switch 121 is overloaded. In addition, the bypass information may indicate a port for bypassing a packet that satisfies a predetermined condition among at least one port of the first network switch 121. [ Bypass to the port allows the processing of packets to be passed on to other network switches. Hereinafter, the packet processing unit 220 of the first network switch 121 may use the bypass information to identify the port to which the packet is to be bypassed and the neighboring network switch.

(vi) Next, according to an embodiment of the present invention, the packet processing unit 220 of the first network switch 121 may receive the packet (S340).

(vii) Next, in accordance with an embodiment of the present invention, the packet processing unit 220 of the first network switch 121 determines whether a packet transmitted to the first network switch 121 after an overload is detected, If so, the transmitted packet can be bypassed to the second network switch 122, which is a neighboring network switch adjacent to the network switch (S350).

According to an embodiment of the present invention, when the packet information is set in the packet processing unit 220 of the first network switch 121 and the packet transmitted to the first network switch 121 meets a predetermined condition, Lt; / RTI > to the second network switch 122, which is a neighboring network switch adjacent to the network switch.

According to one embodiment of the present invention, the predetermined condition described above may mean a state in which the first network switch 121 can not determine the path of the packet itself. The packet processing unit 220 of the first network switch 121 determines that the first network switch 121 can not determine the path of the packet solely for the packet transmitted to the first network switch 121 after the overload is detected , The packet may be bypassed to the second network switch 122, which is a neighboring network switch adjacent to the first network switch 121.

In accordance with one embodiment of the present invention, interaction with other devices other than the first network switch 121 may be required to determine the path of the packet. For example, the determination of the path of the packet may require an upcall to the network controller 110 that controls the first network switch 121. [ For example, the packet processing unit 220 of the first network switch 121 transmits the transmitted packet to the neighboring network switch when there is no flow entry matching the packet transmitted to the flow table of the first network switch 121 You can bypass. If there is no flow entry matched to the packet transmitted to the flow table of the first network switch 121, the packet processing unit 220 of the first network switch 121 transmits the packet to the first network switch 121, It may not be able to determine the port to which the packet is to be transmitted. The packet processing unit 220 of the first network switch 121 is connected to the network controller 110 because the first network switch 121 is overloaded as described above, The processing or arrival of the packet may be delayed, and in some cases, the packet may be corrupted when the upcall to the network controller 110 and the response to the upcall from the network controller 110 are processed. Accordingly, the packet processing unit 220 of the first network switch 121 can reduce the influence of the overload of the first network switch 121 on the processing of the packet by bypassing the transmitted packet directly to the neighboring network switch.

According to an embodiment of the present invention, the data of the packet to be bypassed may be the same as the data of the packet transmitted to the first network switch 121. In addition, the packet processing unit 220 of the first network switch 121 may include the information of the inbound port of the packet in bypassing the packet to the second network switch 122. A packet that is bypassed to a neighboring network switch may include information of the inbound port of the packet. The information of the inbound port can be used to identify an external network or an external device to which the packet is transmitted from another network switch. In other words, in the bypass of the packet, the packet processing unit 220 of the first network switch 121 can keep the data of the packet the same and can add information of the inbound port as information for controlling the packet . For example, data indicating the inbound port may be added to the header of the packet by the packet processing unit 220 of the first network switch 121. [

(viii) Next, according to an embodiment of the present invention, when the packet is bypassed, the packet processing unit 220 of the second network switch 122 may perform processing of the packet (S355). The processing of the packet may include forwarding of the packet and may include an upcall to the network controller 110 for processing of the packet if necessary and a response to the upcall from the network controller 110.

According to an embodiment of the present invention, the packet processing unit 220 of the second network switch 122 may use the information of the inbound port included in the packet in processing the packet. For example, the packet processing unit 220 of the second network switch 122 may use the information of the inbound port to prevent a loop in transmission of the packet. The second network switch 122 is configured to avoid a loop to the inbound port of the first network switch 121 when the path of the packet according to the flow table of the second network switch 122 includes the inbound port included in the packet And may transmit an upcall for the setting of the transmission of the packet to the network controller 110 in order to establish a new path. The network controller 110 may generate configuration information for transmission of the packet by reflecting the bypass in the first network switch 121. [ The setting information may be a flow entry of the flow table. The network controller 110 may generate a packet path in consideration of bypass of a packet or the like and may refer to information of an inbound port of the packet in generating a packet path. In addition, the network controller 110 may generate a flow entry according to the generated route. The flow entry may include information for forwarding the packet according to the generated path. The network controller 110 may send configuration information to the second network switch 122 in response to the upcall.

According to one embodiment of the present invention, the second network switch 122 may be a network switch having the lowest load among the network switches adjacent to the first network switch 121, and may be assumed not to be overloaded. Thus, through the bypass, the processing of the packet is performed in the second network switch 122, so that the delay or loss of the packet can be prevented, and further QoS for the packet can be guaranteed.

According to an embodiment of the present invention, the above-described steps S340 and S350 can be repeatedly performed as an overload of the first network switch 121 is detected and the packet is transmitted to the first network switch 121 have. That is to say, the flow of the flow (or traffic) to the first network switch 121 through the repetition of the above-described steps S340 and S350 is the flow of the uplink to the second network switch 122, which is the neighboring network switch, And the second network switch 122 can perform processing for the bypassed flow.

According to an embodiment of the present invention, when the overload state of the first network switch 121 is released, the steps described below can be performed.

(ix) Next, according to an embodiment of the present invention, the overload detection unit 210 of the first network switch 121 may sense the release of the overload state of the first network switch 121 (S360). When the load of the first network switch 121 falls below the threshold value, the overload detecting unit 210 of the first network switch 121 determines that the overload state of the first network switch 121 is released It can be judged.

In accordance with one embodiment of the present invention, the amount of CPU usage of the network switch, the congestion between the network switch and the network controller, and the number or ratio of flow entries in the flow table of the network switch, etc., Can be referenced. The overload detection unit 210 of the first network switch 121 detects the usage amount of the CPU of the first network switch 121 and the usage amount of the CPU of the first network switch 121, Based on at least one of the number or the ratio of the flow entries in the flow table, the release of the overload state of the first network switch 121 can be detected. For example, the overload detection unit 210 of the first network switch 121 determines that the overload state of the first network switch 121 is released when the usage amount of the CPU of the first network switch 121 is less than a predetermined value can do. If the overload state of the first network switch 121 is released when the congestion between the first network switch 121 and the network controller 110 is below a predetermined level, the overload detection unit 210 of the first network switch 121 . The overload detection unit 210 of the first network switch 121 may determine that the overload state of the first network switch 121 is abnormal when the number or ratio of the flow entries in the flow table of the first network switch 121 is less than a predetermined value It can be judged that it has been released.

(x) Next, according to an embodiment of the present invention, when the release of the overload state is detected, the bypass information for the data path of the first network switch 121 can be deleted. For example, the overload detection unit 210 of the first network switch 121 may delete the bypass information for the data path of the first network switch 121 (S365). Alternatively, the packet processing unit 220 of the first network switch 121 may also perform a set of bypass information. Deletion of the bypass information may indicate that the overload state of the first network switch 121 has been released and that the first network switch 121 has not been overloaded. Due to the deletion of the bypass information, the first network switch 121 can perform the general operation in the processing of the packet.

(xi) Next, according to an embodiment of the present invention, the packet processing unit 220 of the first network switch 121 may receive the packet (S370). The packet received in step S340 may be a packet of the same flow as the packet received in step S340. For example, the packet received in step S340 and the packet received in step S340 may be packets in which the source and destination are the same device or the same network, or the destination is the same device or the same device Network. ≪ / RTI >

(xii) Next, according to an embodiment of the present invention, the packet processing unit 220 of the first network switch 121 detects the release of the overload state, and if the transmitted packet meets a predetermined condition, To the network controller 110 that controls the first network switch 121 (S380). Here, the setting of transmission of a packet may be that the packet is a setup of the corresponding flow, and adding a flow entry matching the packet.

According to one embodiment of the present invention, the predetermined condition described above may mean a state in which the first network switch 121 can not determine the path of the packet alone. For example, the predetermined condition described above may include a condition requiring an upcall to the network controller 110 to control the first network switch 121 for the transmitted packet. For example, the packet processing unit 220 of the first network switch 121 determines whether or not there is a flow entry matched with the packet transmitted to the flow table of the first network switch 121, To the network controller (110).

(xiii) Next, according to an embodiment of the present invention, the network controller 110 may generate setting information for transmission of a packet (S385). The setting information may be a flow entry of the flow table. The network controller 110 may generate an optimal path of a packet in consideration of a network topology or the like, and may refer to information of an inbound port of a packet in generating an optimal path of the packet. In addition, the network controller 110 may generate a flow entry according to the generated optimal route. The flow entry may include information for forwarding the packet according to the generated optimal path. The generated optimal path may include the first network switch 121 in which the overload state is released.

(xiv) Next, according to an embodiment of the present invention, the network controller 110 may transmit setting information on the setting of transmission of the packet to the first network switch 121 (S390).

According to one embodiment of the present invention, the optimal path generated in accordance with the release of the overload state of the first network switch 121 may be different from the path previously defined for the packet. In addition, other network switches other than the first network switch 121 may have a flow entry for the packet that is different from the current optimal path. Thus, the network controller 110 may transmit the generated configuration information to a plurality of network switches, such as the second network switch 122, if the generated configuration information differs from the previous path or flow entry for the packet .

(xv) Next, according to an embodiment of the present invention, the packet processing unit 220 of the first network switch 121 may perform the processing of the packet using the setting information (S395).

According to one embodiment of the invention, the configuration information may comprise a flow entry for the packet. That is to say, the transmitted flow entry can be matched to the packet, and the packet processing unit 220 of the first network switch 121 can perform the processing of the packet, such as packet forwarding, according to the action of the flow entry.

The embodiments of the present invention described above can be implemented in the form of program instructions that can be executed through various computer components and recorded on a computer-readable recording medium. The computer-readable recording medium may include program commands, data files, data structures, and the like, alone or in combination. The program instructions recorded on the computer-readable recording medium may be those specially designed and constructed for the present invention or may be those known to those skilled in the art of computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks and magnetic tape, optical recording media such as CD-ROMs and DVDs, magneto-optical media such as floptical disks, media, and hardware devices specifically configured to store and execute program instructions such as ROM, RAM, flash memory, and the like. Examples of program instructions include machine language code such as those generated by a compiler, as well as high-level language code that can be executed by a computer using an interpreter or the like. The hardware device may be configured to operate as one or more software modules for performing the processing according to the present invention, and vice versa.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed exemplary embodiments, but, on the contrary, Those skilled in the art will appreciate that various modifications, additions and substitutions are possible, without departing from the scope and spirit of the invention as disclosed in the accompanying claims.

Therefore, the spirit of the present invention should not be construed as being limited to the above-described embodiments, and all of the equivalents or equivalents of the claims, as well as the following claims, I will say.

110: Network controller
121: first network switch
122: second network switch
123: third network switch
124: fourth network switch
200: Network switch
210: Overload detection unit
220:
230:

Claims (20)

A packet processing method of a network switch,
(a) detecting whether the network switch is overloaded; And
(b) if the network switch is detected as overloaded and the network switch is unable to determine solely the path of a predetermined packet transmitted to the network switch, the packet is transmitted to a neighboring network switch By-
Lt; / RTI >
The step (b)
When there is no flow entry matched with the transmitted predetermined packet in the flow table of the network switch in a state where the network switch is detected as being overloaded, the network switch transmits, to the network controller controlling the network switch, Bypassing the transmitted predetermined packet to the neighboring network switch, the neighboring network switch requests the uplink directly to the network controller, confirms the path from the network controller, And the packet is transmitted to the destination. delete delete The method according to claim 1,
In the step (a)
Wherein the network switch detects the overload based on at least one of a usage amount of the central processing unit of the network switch, a congestion between the network switch and a network controller, and a number or ratio of flow entries in the flow table of the network switch. Lt; / RTI > delete The method according to claim 1,
Wherein the neighboring network switch is selected based on at least one of a traffic statistic value of each network switch of the plurality of network switches, a usage amount of the central processing unit, and a number or ratio of flow entries in the flow table. The method according to claim 1,
In the step (b)
Wherein a packet bypassed to the neighboring network switch includes information of an inbound port of the packet. The method according to claim 1,
After the step (b)
(c) the network switch detecting a release of the overload state;
(d) after the network switch detects the release of the overload state, transmitting an upcall for setting a path of a packet to a network controller controlling the network switch; And
(e) receiving, by the network switch, information on the setting of the path of the packet from the network controller
≪ / RTI > A computer-readable recording medium having recorded thereon a computer program for executing the method according to any one of claims 1, 4, 6 to 8. A network switch comprising:
An overload detector for detecting an overload of the network switch; And
If it is determined that the network switch is overloaded and the network switch is unable to independently determine a path of a predetermined packet transmitted to the network switch, the packet is bypassed to a neighboring network switch adjacent to the network switch, a packet processing unit for
Lt; / RTI >
When there is no flow entry matched with the transmitted predetermined packet in the flow table of the network switch while the network switch is detected as being overloaded,
The packet processing unit,
Bypassing the transmitted predetermined packet to the neighboring network switch without requiring an upcall to the network controller controlling the network switch so that the neighboring network switch requests the network controller to upcall directly, And to transmit the packet through the determined path. delete delete 11. The method of claim 10,
The overload detection unit includes:
Wherein the network switch detects the overload based on at least one of a usage amount of the central processing unit of the network switch, a congestion between the network switch and a network controller, and a number or ratio of flow entries in the flow table of the network switch. delete 11. The method of claim 10,
Wherein the neighboring network switch comprises:
Wherein the network switch is selected based on at least one of a traffic statistic value of each network switch of the plurality of network switches, a usage amount of the central processing unit, and a number or ratio of flow entries in the flow table. 11. The method of claim 10,
Wherein a packet bypassed to the neighboring network switch includes information of an inbound port of the packet. 11. The method of claim 10,
The overload detection unit includes:
Detects the release of the overload state,
The packet processing unit,
Wherein the controller detects an overload state and transmits an upcall for setting a path of a packet to a network controller that controls the network switch and receives information on a path setting of the packet from the network controller, Network switch. A plurality of network switches, and
Network controller
Lt; / RTI >
Wherein the first network switch of the plurality of network switches senses whether the first network switch is overloaded and, in a state where the first network switch is detected as being overloaded, When there is no flow entry matched with a predetermined packet transmitted to the network switch, the first network switch does not request the network controller that controls the first network switch, 2 network switch to allow the second network switch to directly request the uplink to the network controller to confirm the path from the network controller and to transmit the packet over the determined path. . delete 19. The method of claim 18,
When the first network switch detects the release of the overload state, the first network switch transmits to the network controller an upcall for the setting of the path of the packet, and sets the path of the packet from the network controller Lt; RTI ID = 0.0 > 1, < / RTI >

Images