KR20170057770A - Method for controlling transsion of packet in virtual switch - Google Patents

Abstract

The present invention relates to a method for controlling packet transmission performed by a virtual switch. The method comprises: a step of receiving, from a network interface card, hardware switching port information which supports packet transmission between virtual machines; a step of transmitting, to a software defined networking (SDN) controller, the received hardware switching port information and software switching port information for the packet transmission between the virtual machines; a step of receiving a packet transmission rule from the SDN controller; and a step of transmitting the received packet transmission rule to the network interface card. The packet transmission rule can determine hardware switching based on traffic priority information, or hardware switching through the hardware switching port for a specific flow according to a policy on distribution of software switching. Therefore, an SDN controller outside a server can control switching of a virtual switch and a hardware off-load function by the network interface card, and a hardware off-load function can be dynamically changed or managed by the outside of the server.

Description

TECHNICAL FIELD [0001] The present invention relates to a packet transfer control method for a virtual switch,

The present invention relates to a packet transmission control method for a virtual switch, and more particularly, to a packet transmission control method performed by a virtual switch supporting communication between virtual machines.

As is known, a virtual switch connects a plurality of virtual machines existing in one physical server, and connects a virtual machine and a network interface card.

However, since such a virtual switch is driven by software, there is performance deterioration as compared with a hardware switch. Particularly, traffic to be transmitted between virtual machines increases, and deterioration of throughput performance is serious when small size packets are frequently transmitted. In addition, CPU resources are occupied for packet processing and transmission. This has the disadvantage of wasting CPU resources to be used for processing the application.

In order to overcome the drawbacks of such a virtual switch, a hardware VEB (Virtual Ethernet Bridge) technology has been used. Hardware VEB is a technology that offloads the switching function of a virtual switch to a physical network interface card. Using Virtual Role I / O Virtualization (SR-IOV) technology, virtual machines and network interface cards can be directly connected to transmit packets, enabling Layer 2 (L2) switching on the network interface card via hardware VEB technology, Lt; RTI ID = 0.0 > a < / RTI >

However, this conventional technology can not dynamically change or manage the hardware offload function by the network interface card outside the server, and can not use the hardware VEB at the same time with the virtual switch. In addition, there is a problem that the capacity that can be switched through the hardware VEB is limited depending on the table size of the network interface card and the number of virtual functions (VF).

Korean Patent Publication No. 10-2013-0039626, published on April 22, 2013.

DISCLOSURE Technical Problem The present invention has been proposed in order to solve the problems of the related art as described above, and it is an object of the present invention to provide a system and method for switching a virtual switch and a hardware offload function by a network interface card, And a packet transmission control method performed by the switch.

The problems to be solved by the present invention are not limited to those mentioned above, and another problem to be solved can be clearly understood by those skilled in the art from the following description.

According to an aspect of the present invention, there is provided a packet transfer control method performed by a virtual switch, the method comprising: receiving hardware switching port information for supporting packet transmission between virtual machines from a network interface card; And transmitting software switching port information for packet transmission between the virtual machines to an SDN controller, receiving a packet transmission rule from the SDN controller, transmitting the received packet transmission rule to the network interface card Wherein the packet forwarding rule can determine hardware switching over the hardware switching port for a particular flow according to hardware switching based on traffic priority information and a distribution policy of software switching.

According to the embodiment of the present invention, the switching of the virtual switch and the hardware offload function by the network interface card can be controlled by the SDN controller outside the server, and the hardware offload function can be dynamically changed can do. Therefore, packets of high priority traffic can be dynamically allocated to hardware switches to accelerate packet transmission speed. In addition, because the hardware switch takes care of the switching that the software switch needs to process using the CPU, more CPU resources can be allocated to applications running in the virtual machine.

1 is a system configuration diagram of a compute node and a SDN controller including a virtual switch for performing a packet transmission control method according to an embodiment of the present invention.
2 is a flowchart illustrating a packet transmission control method performed by a virtual switch according to an embodiment of the present invention.
3 is a flowchart illustrating a packet transmission process performed by a network interface card according to an embodiment of the present invention.

BRIEF DESCRIPTION OF THE DRAWINGS The advantages and features of the present invention and the manner of achieving them will become apparent with reference to the embodiments described in detail below with reference to the accompanying drawings. The present invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art. To fully disclose the scope of the invention to those skilled in the art, and the invention is only defined by the scope of the claims.

In the following description of the present invention, a 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. The following terms are defined in consideration of the functions in the embodiments of the present invention, which may vary depending on the intention of the user, the intention or the custom of the operator. Therefore, the definition should be based on the contents throughout this specification.

Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings.

1 is a system configuration diagram of a compute node and a SDN controller including a virtual switch for performing a packet transmission control method according to an embodiment of the present invention.

As shown, the compute node 100 includes a guest OS 110, a host OS 120, and a network interface card 130.

The guest OS 110 includes a plurality of virtual machines 111, 112, 113, Although a total of four virtual machines are shown in FIG. 1, the number of virtual machines is only an example, and there is no limitation within the capacity provided by the server performance and the network interface card.

The host OS 120 includes a virtual switch 121 and the virtual switch 121 includes a second offload driver 121a.

The network interface card 130 includes a hardware offload module 131 and the hardware offload module 131 includes a virtual Ethernet bridge 131a.

The SDN controller 200 includes a traffic classifier 210, a hardware offload manager 220, and a first offload driver 230.

A plurality of APIs are defined for the interface 10 between the SDN controller 200 and the virtual switch 121 and the interface 20 between the virtual switch 121 and the network interface card 130.

In the network thus configured, the SDN controller 200 determines whether the virtual switch 121 connects between the virtual machines 111, 112, 113, and 114 included in the compute node 100 through the software switching 30, And provides a packet transmission rule used when the Ethernet bridge 131a determines whether to connect through the hardware switching 40. [

The virtual switch 121 of the computer node 100 connects the connection between the virtual machines 111, 112, 113 and 114 via the software switching 30 and transmits the packet transmission rule provided from the SDN controller 200 To the network interface card (130).

The virtual Ethernet bridge 131a of the computer node 100 connects the connection between the virtual machines 111, 112, 113 and 114 via the hardware switching 40. [ That is, the switching connection function of the virtual switch 121 is offloaded.

The traffic classifier 210 of the SDN controller 200 stores traffic priority information for packet transmission according to priority information defined for subscribers, application priority information defined for each subscriber, and flow characteristics .

The hardware offload manager 220 of the SDN controller 200 controls the hardware offload module 131 included in the compute node 100 to offload some functions of the virtual switch 121, And manages them dynamically. To this end, the hardware offload manager 220 provides a packet transmission rule to the network interface card 130 through the first offload driver 230 and the second offload driver 121a. In addition, the hardware offload manager 220 transmits a hardware offload activation signal for the function occupying the CPU resources of the compute node to the network interface card 210 via the first offload driver 230 and the second offload driver 121a. (130).

The virtual switch 121 included in the compute node 100 receives information of a hardware switching port that supports packet transmission between the network interface card 130 and the virtual machines 111, 112, 113, and 114, Information of the software switching port for packet transmission between the virtual machines and information of the received hardware switching port to the SDN controller 200.

The virtual switch 121 also receives a packet transmission rule from the SDN controller 200 that can determine hardware switching through a hardware switching port for a specific flow according to hardware switching based on traffic priority information and distribution policy of software switching And transmits the received packet transmission rule to the network interface card 130.

2 is a flowchart illustrating a packet transmission control method performed by a virtual switch according to an embodiment of the present invention.

As shown, the method includes receiving (303,305) information of a hardware switching port that supports packet transmission between virtual machines from a network interface card after a new flow arrives at a hardware offload module .

Then, the step of transmitting (307) the information of the software switching port for packet transmission between the virtual machines and the information of the received hardware switching port to the SDN controller.

Then, according to the hardware switching based on the traffic priority information and the distribution policy of the software switching in a state where the traffic priority policy is acquired by the traffic classifier, a packet transmission rule that can determine the hardware switching through the hardware switching port (301, 309, 311, 313) from the SDN controller.

Next, the method includes transmitting the received packet transmission rule to the network interface card so that hardware switching or software switching is performed according to the packet transmission rule (315, 317, 319, 321).

3 is a flowchart illustrating a packet transmission process performed by a network interface card according to an embodiment of the present invention.

As shown, if the arrived packet is a new flow that is not included in the existing packet transmission rule, it includes steps 401, 403, and 405 for updating the packet transmission rules through cooperation with the SDN controller.

If it is determined that the received packet is included in the existing packet transmission rule, step (S407, 409) is performed to determine whether the packet is a hardware switching object based on the packet transmission rule.

Then, if the object is a hardware switching object, transmitting (411) a packet to the virtual machine through a hardware switching port.

Hereinafter, a packet transmission process in a compute node including a virtual switch for performing a packet transmission control method according to an embodiment of the present invention will be described in more detail with reference to FIG. 1 to FIG.

First, for the interface 10 between the SDN controller 200 and the virtual switch 121 and the interface 20 between the virtual switch 121 and the network interface card 130, a function activation API, a port configuration API, An Ethernet bridge API is defined. The function activation API is an API for controlling a hardware offload activation signal for occupying the CPU resources of the compute node 100 such as VXLAN tunneling, segmentation, and checksum checking. The port configuration API is an API for managing the connection configuration of VF (Virtual Function) and PF (Physical Function) of the network interface card 130. The hardware virtual Ethernet bridge API is an API for reflecting the packet transmission rule for switching to the flow table of the network interface card 130.

The traffic classifier 210 stores traffic priority information for packet transmission according to priority information for tenants, application priority information defined for each subscriber, and priority information defined according to flow characteristics. The priorities of subscribers can be defined differently, and traffic priority according to application priority can be defined for each subscriber. In addition, a priority order can be defined according to the flow characteristics. For example, since the elephant flow increases the transmission delay of the mice flow, a high priority is given to the selected elephant flow to increase the transmission speed. In this case, regardless of hardware switching or software switching, dedicated switching You can create and schedule lanes. Internet traffic statistics show that a small number of elephant flows generally account for about 80% of the total traffic volume and a large number of meal flows have a heavy-tailed distribution that accounts for the remaining traffic volume. A long-lived flow may be referred to as an " eliptant flow ", and a bursty flow or a short-lived flow may be referred to as a " meal flow ". For example, if the session duration is 15 minutes or longer, it can be classified as an element flow, and if the specific flow occupies 10% or more of the bandwidth of the link utilization, it can be classified as an element flow.

In the initial setup of the system, the traffic classifier 210 inputs and obtains a priority ranking rule by a user or an administrator in advance (301).

When a new flow not included in the existing packet transmission rule arrives at the network interface card 130, the hardware offload module 131 transmits the information of the hardware switching port of the virtual Ethernet bridge 131a to the virtual switch 121 (305).

Subsequently, the virtual switch 121 transmits the information of the hardware switching port of the virtual Ethernet bridge 131a to the hardware offload manager 220 through the second offload driver 121a and the first offload driver 230 . Here, the virtual switch 121 conveys the information of the software switching port for packet transmission between the virtual machines 111, 112, 113, and 114 to the hardware offload manager 220 together (307).

Then, the hardware offload manager 220 receives the traffic priority information from the traffic classifier 210 (309), and based on the hardware switching port information and the traffic priority information, (311) a packet transmission rule that can determine hardware switching through a switching port.

The generated packet transmission rule is transmitted to the virtual switch 121 through the first offload driver 230 and the second offload driver 121a 313 and the virtual switch 1210 transmits the packet transmission rule to the network interface card To the hardware offload module 131 of the controller 130 (315).

Here, the hardware offload module 131 updates the existing packet transmission rule to the packet transmission rule received from the virtual switch 121 (317), performs hardware switching according to the updated packet transmission rule, 112, 113, and 114 (319).

If the packet flow is not a hardware switching object according to the packet transmission rule updated in step 317, the hardware offload module 131 may transmit the virtual machines 111, 112, 113 , 114 (321).

In addition, the hardware offload manager 220 of the SDN controller 200 transmits the VXLAN tunneling, segmentation, checksum check, etc. to the virtual switch 121 through the first offload driver 230 and the second offload driver 121a It can transmit a hardware offload enable signal for a function that occupies the CPU resources of the compute node 100 and is transmitted to the hardware offload module 131 in the network interface card 130 by the virtual switch 121 . Then, a tunneling function for forming a VXLAN packet by attaching a header before an Ethernet packet, a segmentation function for cutting a small size packet when a large packet is physically transmitted, and a checksum check function for a VXLAN packet are provided in a hardware offload module 131 ), And perform and process it.

As described above, according to the embodiment of the present invention, the SDN controller outside the server can control the switching of the virtual switch and the hardware offload function by the network interface card, and the hardware offload function can be dynamically changed Or management. Therefore, the packet of high priority traffic can be dynamically allocated to the hardware switch to accelerate the transmission speed. This allows the hardware switch to do the switching that the software switch needs to process using the CPU, so that more CPU resources can be allocated to applications running in the virtual machine.

Each block of the block diagrams attached hereto and combinations of steps of the flowchart diagrams may be performed by computer program instructions. These computer program instructions may be embedded in a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus so that the instructions, which may be executed by a processor of a computer or other programmable data processing apparatus, And means for performing the functions described in each step are created. These computer program instructions may also be stored in a computer usable or computer readable memory capable of directing a computer or other programmable data processing apparatus to implement the functionality in a particular manner so that the computer usable or computer readable memory It is also possible for the instructions stored in the block diagram to produce a manufacturing item containing instruction means for performing the functions described in each block or flowchart of the block diagram. Computer program instructions may also be stored on a computer or other programmable data processing equipment so that a series of operating steps may be performed on a computer or other programmable data processing equipment to create a computer- It is also possible that the instructions that perform the processing equipment provide the steps for executing the functions described in each block of the block diagram and at each step of the flowchart.

Also, each block or each step may represent a module, segment, or portion of code that includes one or more executable instructions for executing the specified logical function (s). It should also be noted that in some alternative embodiments, the functions mentioned in the blocks or steps may occur out of order. For example, two blocks or steps shown in succession may in fact be performed substantially concurrently, or the blocks or steps may sometimes be performed in reverse order according to the corresponding function.

The foregoing description is merely illustrative of the technical idea of the present invention, and various changes and modifications may be made by those skilled in the art without departing from the essential characteristics of the present invention. Therefore, the embodiments disclosed in the present invention are intended to illustrate rather than limit the scope of the present invention, and the scope of the technical idea of the present invention is not limited by these embodiments. The scope of protection of the present invention should be construed according to the following claims, and all technical ideas within the scope of equivalents should be construed as falling within the scope of the present invention.

According to the embodiment of the present invention, the SDN controller outside the server can control the switching of the virtual switch and the hardware offload function by the network interface card, and the hardware offload function can be dynamically changed or managed outside the server .

The present invention can be used in various systems including a virtual node connecting a plurality of virtual machines existing in a physical server and connecting a virtual machine and a network interface card.

100: compute node 111, 112, 113, 114: virtual machine
121: virtual switch 130: network interface card
131: hardware offload module 131a: virtual Ethernet bridge
200: SDN controller 210: traffic classifier
220: Hardware offload manager

Claims (6)

Receiving hardware switching port information to support packet transmission between virtual machines from a network interface card;
Sending information of the received hardware switching port and software switching port information for packet transmission between the virtual machines to a Software-Defined Networking (SDN) controller;
Receiving a packet transmission rule from the SDN controller;
And transmitting the received packet transmission rule to the network interface card,
Wherein the packet transfer rules are capable of determining hardware switching over the hardware switching port for a particular flow according to hardware switching based on traffic priority information and a distribution policy of software switching
A method for controlling packet transmission performed by a virtual switch. The method according to claim 1,
The traffic priority information includes priority information for subscribers or application priority information defined for each subscriber
A method for controlling packet transmission performed by a virtual switch. The method according to claim 1,
The traffic priority information includes priority information defined according to flow characteristics
A method for controlling packet transmission performed by a virtual switch. The method of claim 3,
Wherein the specific flow selected in accordance with the flow characteristic is switched to any one of the hardware switching and the software switching
A method for controlling packet transmission performed by a virtual switch. The method according to claim 1,
Further comprising receiving from the SDN controller a hardware offload enable signal for a function that occupies CPU resources of the compute node including the virtual machine and the virtual switch and forwarding the hardware offload enable signal to the network interface card
A method for controlling packet transmission performed by a virtual switch. In order for the processor to perform the packet transmission control method of any one of claims 1 to 5
A computer program stored on a computer readable recording medium.

Images