KR102481033B1 - Method for client device registration and ultra-low latency large network packet management - Google Patents

Abstract

According to an embodiment of the present invention, a method for client device registration and ultra-low latency large network packet management comprises the steps of: (a) registering a plurality of client devices for each of host devices composed of a plurality of slave servers and a master server; (b) transmitting, by the client device, client role information and network packet type information to be processed to the host device, and sharing, by the host device, the client role information and the network packet type information with other host devices; (c) sharing, by the host device, the client role information among information received from the other host device with a client device managed by the host device; and (d) when the client device transmits a specific packet to the host device along with the network packet type information, referring to the previously stored network packet type information for each client device, and transmitting, by the host device, the specific packet to the other host device connected to the client device capable of processing the specific packet, such that the specific packet can be processed in the other client device.

Description

Method for managing large network packets for client device registration and ultra-low latency

The present invention relates to a method for registering a client device and managing large-capacity network packets for ultra-low latency, and more particularly, when a client device registered in a host device composed of a plurality of servers forwards a network packet to a client device of another host device. , it relates to a method for minimizing the delay that occurs.

Nowadays, a variety of information is transmitted through a network, and a server device is essential in order to share such a variety of information.

However, if a large number of devices access the server device for this purpose and countless packets are transmitted through the server device, the server becomes overloaded. Therefore, traffic is concentrated on one server, and eventually the amount of traffic is exceeded and a problem may occur in the server.

In order to solve this problem, the conventional technology distributes several physical servers to control the amount of network traffic that one server has to handle. However, if a common network packet is distributed to multiple physical servers, it is simply possible to distribute the packets to multiple servers, but a specific server does not have to process packets, and a specific server needs to process packets. A targeting function is also required.

In this packet distribution process, there are network delays that may occur between packet movements and packet distribution efficiency problems for several randomly distributed clustered servers.

In order to solve this problem, servers divided into several physical servers are grouped into one clustering structure, clustering is automated or flexible expansion is possible, and the delay problem between network packet transmissions occurring between clustered servers is minimized. There is a need for a packet processing technology that does not cause problems in maintaining real-time performance by minimizing delay even when large-capacity network packets such as communication occur.

The present invention is to solve the above-mentioned problems of the prior art, and the delay between network packet transmissions occurring between clustering servers that binds servers divided into several physical servers into one clustering structure and enables automation or flexible expansion. The purpose is a packet processing technology that minimizes problems and minimizes delay even when large-capacity network packets occur.

However, the technical problem to be achieved by the present embodiment is not limited to the technical problem as described above, and other technical problems may exist.

As a technical means for achieving the above technical problem, in the method for registering a client device and managing large-capacity network packets for ultra-low delay according to an embodiment of the present invention, (a) a host device composed of a plurality of slave servers and a master server Registering a plurality of client devices for each of the above; (b) the client device transmitting client role information and network packet type information to be processed to the host device, and the host device sharing the client role information and network packet type information with other host devices; (c) sharing, by the host device, client role information, among information received from other host devices, with a client device managed by the host device; and (d) when the client device forwards a specific packet to the host device along with network packet type information, the host device refers to previously stored network packet type information for each client device, and processes the specific packet. and transferring the specific packet to another host device connected to the device so that the specific packet can be processed by another client device.

Also, devices managed by the host device correspond to any one of the clustered host devices and the client devices that delivered the authentication packet to the host device.

In the step (b), the client device additionally transfers client role information and network packet type information when the client role information and network packet type information to be processed are changed.

Also, prior to the step (a), (a-1) the first slave server and the second slave server transmitting access and authentication packets to the master server; (a-2) transmitting, by the master server, a cluster number and a list of other accessible slave servers to the first slave server and the second slave server; and (a-3) when the first slave server accesses the master server earlier than the second slave server, the second slave server transmits an access and authentication packet to the first slave server, and the first Slave server accepting access; Further, the cluster number is issued sequentially according to the order in which the master server transfers the access and authentication packets of the slave server, and in connection between the slave servers, the master server of the slave server The slave server at the earliest delivery time of the access and authentication packet becomes the connection acceptor, and the slave server at the later time at the delivery time of the access and authentication packet of the slave server to the master server becomes the accessor role.

In addition, the step (a) may include transmitting an authentication packet after the client device accesses any one of the clustered host devices composed of a plurality of slave servers and master servers; and registering, by the host device receiving the authentication packet, the client device as one of client devices managed by the host device.

Also, the network packet type information is shared among the host devices, but not shared among the client devices, and the host device transmits the client role information when a new client device accesses the clustering server.

In addition, after step (d), the step of storing, by the 1-1st client device registered in the 1st slave server, the role of the 2-1st client device registered in the 2nd slave server; notifying, by the second slave server, the end of driving of the 2-1 client device to the first slave server when the driving of the 2-1 client device is terminated; and the 1-1 client device removes management information from the stored client device list according to the end of driving of the 2-1 client device, and refers to the previously stored client role information to select other client devices capable of the same role. It includes; finding and requesting communication.

In addition, each host device and each client device may include a main thread; and a network packet processing thread including a receiving thread and a transmitting thread.

In addition, a payload packet is a packet that must be generated in a client device and transmitted to be processed by another client device, and a control packet is a packet specifying a path of the payload packet, and each host device and each client device, When a control packet is received, it is processed through the main thread, and when a payload packet is received, it is processed through the network packet processing thread.

In addition, each server includes a network packet processing thread for each of the other servers, and any one server processes a network packet processing thread (including a sending thread and a receiving thread) for the first slave server and a network packet processing thread for the second slave server. When a processing thread (including a sending thread and a receiving thread) is included, when any one of the above servers receives a packet from the first slave server, the payload packet is received through the receiving thread for the first slave server, and control When the packet indicates transmission to the second slave server, the payload packet is transferred to the second slave server through a transmission thread for the second slave server.

In addition, the client device or the host device divides the entire packet into frames in units of MTU and transmits the entire packet to the network for integrity of the packet, and the packet includes a frame number, a start of a packet, an end of a packet, and a size of the frame. and the MTU size does not exceed 1,500 bytes.

In addition, information containing identification information indicating whether the corresponding packet is a control packet or a payload packet is placed at the front of the corresponding packet.

In addition, the client device or the host device receives the packet and identifies whether it is a control packet or a payload packet by analyzing only the first structure of the packet, and determines which thread to process between the main thread and the network packet processing thread. do.

A system for managing large-capacity network packets for client device registration and ultra-low latency, comprising a plurality of host devices and a plurality of client devices, wherein the devices execute a method for performing large-capacity network packet management, the method comprising (a) registering a plurality of client devices for each of the host devices composed of a plurality of slave servers and a master server; (b) the client device transmitting client role information and network packet type information to be processed to the host device, and the host device sharing the client role information and network packet type information with other host devices; (c) sharing, by the host device, client role information, among information received from other host devices, with a client device managed by the host device; and (d) when the client device forwards a specific packet to the host device along with network packet type information, the host device refers to previously stored network packet type information for each client device, and processes the specific packet. It is forwarded to another host device connected to the device so that the specific packet can be processed by another client device.

The present invention is to solve the above-mentioned problems of the prior art, and to minimize the delay problem between network packet transmissions occurring between host devices composed of a plurality of servers, and to process packets that minimize delay even when large-capacity network packets occur There are advantages.

1 is a structural diagram of a host device according to an embodiment of the present invention.
2 is a flowchart of a process in which a plurality of physical servers become host devices by clustering according to an embodiment of the present invention.
FIG. 3A is a flowchart illustrating a process of restoring a master server of a host device according to an embodiment of the present invention when a disruption occurs.
FIG. 3B is a flowchart illustrating a process of restoring when a second slave server of a host apparatus is interrupted according to an embodiment of the present invention.
4A to 4C are exemplary diagrams illustrating a process in which a plurality of physical servers become host devices through clustering according to an embodiment of the present invention.
5A to 5D are exemplary diagrams illustrating a restoration process when a interruption occurs in a master server or a second slave server of a host device according to an embodiment of the present invention.
6 is a flowchart illustrating a process in which a client device transmits a packet through a host device according to an embodiment of the present invention.
7 is a structural diagram illustrating a plurality of client devices connected to a host device according to an embodiment of the present invention.
8 is an exemplary diagram illustrating network packet control through thread separation according to an embodiment of the present invention.
9 is an exemplary diagram illustrating a configuration of an MTU according to an embodiment of the present invention.
10 is an exemplary diagram illustrating packet delivery delay minimization according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail so that those skilled in the art can easily practice the present invention with reference to the accompanying drawings. However, the present invention may be embodied in many different forms and is not limited to the embodiments described herein. And in order to clearly explain the present invention in the drawings, parts irrelevant to the description are omitted, and similar reference numerals are attached to similar parts throughout the specification.

Throughout the specification, when a part is said to be "connected" to another part, this includes not only the case where it is "directly connected" but also the case where it is "electrically connected" with another element interposed therebetween. . In addition, when a certain component is said to "include", this means that it may further include other components without excluding other components unless otherwise stated.

In this specification, a "unit" includes a unit realized by hardware, a unit realized by software, and a unit realized using both. Further, one unit may be realized using two or more hardware, and two or more units may be realized by one hardware. On the other hand, '~ unit' is not limited to software or hardware, and '~ unit' may be configured to be in an addressable storage medium or configured to reproduce one or more processors. Therefore, as an example, '~unit' refers to components such as software components, object-oriented software components, class components, and task components, processes, functions, properties, and procedures. , subroutines, segments of program code, drivers, firmware, microcode, circuitry, data, databases, data structures, tables, arrays and variables. Functions provided within components and '~units' may be combined into smaller numbers of components and '~units' or further separated into additional components and '~units'. In addition, components and '~units' may be implemented to play one or more CPUs in a device or a secure multimedia card.

A “user terminal” referred to below may be implemented as a computer or portable terminal capable of accessing a server or other terminals through a network. Here, the computer is, for example, a laptop, desktop, laptop, VR HMD (for example, HTC VIVE, Oculus Rift, GearVR, DayDream, PSVR, etc.) equipped with a web browser, etc. can include Here, the VR HMD is for PC (for example, HTC VIVE, Oculus Rift, FOVE, Deepon, etc.), for mobile (for example, GearVR, DayDream, Stormtrooper, Google Cardboard, etc.), and for console (PSVR) and It includes all independently implemented Stand Alone models (eg, Deepon, PICO, etc.). A portable terminal is, for example, a wireless communication device that ensures portability and mobility, and includes not only a smart phone, tablet PC, and wearable device, but also Bluetooth (BLE, Bluetooth Low Energy), NFC, RFID, and ultrasonic waves (Ultrasonic). , various devices equipped with communication modules such as infrared, Wi-Fi, and LiFi. In addition, "network" refers to a connection structure capable of exchanging information between nodes such as terminals and servers, such as a local area network (LAN), a wide area network (WAN), and the Internet. (WWW: World Wide Web), including wired and wireless data communications networks, telephone networks, and wired and wireless television communications networks. Examples of wireless data communication networks include 3G, 4G, 5G, 3rd Generation Partnership Project (3GPP), Long Term Evolution (LTE), World Interoperability for Microwave Access (WIMAX), Wi-Fi, Bluetooth communication, infrared communication, ultrasonic communication, visible light communication (VLC: Visible Light Communication), LiFi, and the like, but are not limited thereto.

First, FIG. 1 is a structural diagram of a host device according to an embodiment of the present invention.

Referring to FIG. 1, the host device of the present invention includes a master server 100, a first slave server 110, a second slave server, and a third slave server.

The master server 100 manages access of a plurality of slave servers and delivers a list of managed slave servers to the connected slave servers. Also, cluster numbers are issued to slave servers.

The slave server is a device excluding the master server 100 among a plurality of servers, and first accesses the master server 100, receives a list of other slave servers to be connected, and tries to connect to the other slave servers.

Hereinafter, a clustering process of a plurality of physical servers according to an embodiment of the present invention will be described in detail with reference to FIG. 2 .

First, the first slave server 110 transmits an access and authentication packet to the master server 100 (S201). The master server 100 sends the first slave server 110 a cluster number and a list of other accessible slave servers. is delivered. (S202)

The above steps S201 and S202 will be described with reference to FIG. 4A. When the server is running and only the master server 100 and the first slave server 110 exist, the first slave server 110 accesses the master server 100 and sends a packet authenticating that it is the first slave server 110. convey The master server 100 checks the connection of the first slave server 110 and transmits a unique first cluster number and a list of other accessible slave servers. At this time, since other slave servers other than the first slave server 110 do not exist, the list of other slave servers may be empty.

The cluster number mentioned above is a unique number sequentially issued so that the master server 100 does not overlap with the order in which the connection and authentication packets of the slave servers are transmitted. Depending on the cluster number, it plays the role of connection acceptor and accessor between each slave server. At this time, the slave server at the earliest when the slave server connects to the master server 100 and delivers the authentication packet becomes the connection acceptor, and the slave server connects to the master server 100 and delivers the authentication packet later. The server becomes the accessor.

The second slave server transfers access and authentication packets to the master server 100. (S203) The master server 100 transfers the cluster number and a list of other accessible slave servers to the second slave server. (S204)

Steps S203 to S204 will be described below with reference to FIG. 4B. If the second slave server additionally accesses after step S202, the second slave server connects to the master server 100 and transmits a packet verifying that it is the second slave server. The master server 100 checks the connection of the second slave server and transmits a unique second cluster number and a list of other accessible slave servers. At this time, since the first slave server 110 is connected to the master server 100 as a visitor, the first slave server 110 may exist in the list of other slave servers.

The second slave server that accesses the master server 100 in the second priority transfers the access and authentication packet to the first slave server 110, and the first slave server 110 accepts the connection. (S205)

Process S205 will be described with reference to FIG. 4B below. In step S204, the second slave server receives a list of other slave servers from the master server 100, and the previously connected first slave server 110 is included in the corresponding list. Accordingly, the second slave server accesses the first slave server 110 and transmits a packet verifying that it is the second slave server. The first slave server 110 checks the second slave server and accepts the connection. Thus, the first slave server 110 becomes an access acceptor, and the second slave server becomes an accessor.

The 3rd slave server transfers the access and authentication packet to the master server 100, and the master server 100 transfers the cluster number and the list of other accessible slave servers to the 3rd slave server. (S206)

Process S206 will be described below with reference to FIG. 4C. Until step S205, the first slave server 110 and the second slave server were issued a unique cluster number through the master server 100 and clustered. Then, when the third slave server accesses, the third slave server connects to the master server 100 and transmits a packet verifying that it is the third slave server. The master server 100 checks the connection of the third slave server and transmits a unique third cluster number and a list of other accessible slave servers. At this time, since the first slave server 110 and the second slave server are connected to the master server 100 as visitors, the first slave server 110 and the second slave server may exist in the list of other slave servers. there is.

The third slave server, which accesses the master server 100 in the second order, transfers the connection and authentication packet to the first slave server 110 and the second slave server, and the first slave server 110 and the second slave server Access is accepted. (S207)

The process of S207 will be described below with reference to FIG. 4C. In step S206, the third slave server receives a list of other slave servers from the master server 100, and the list includes the first slave server 110 and the second slave server that were already connected. Accordingly, the third slave server connects to the first slave server 110 and the second slave server and transmits a packet verifying that it is the third slave server. The first slave server 110 and the second slave server check the third slave server and accept the connection. Thus, the first slave server 110 and the second slave server become connection acceptors, and the third slave server becomes an accessor for the first slave server 110 and the second slave server.

Hereinafter, with reference to FIG. 3A, a restoration process when a shutdown occurs in the master server 100 according to an embodiment of the present invention will be described in detail.

First, the master server 100 is disconnected. (S301) The connection to the new slave server is stopped until the master server 100 connection is restored. (S302) The previously connected slave servers try to reconnect until the master server 100 is restored. .(S303) When the connection of the master server 100 is restored, each slave server transmits authentication packets in the order of connection. (S304) The master server 100 transmits the updated new cluster number to each slave server. (S305)

Steps S301 to S305 will be described with reference to FIGS. 4C and 5A.

4C is an exemplary diagram illustrating a clustering server including a master server 100, a first slave server 110, a second slave server, and a third slave server. Each slave server receives the first cluster number, the second cluster number, and the third cluster number from the master server 100 . In addition, each slave server is connected to the master server 100 as a visitor, the second slave server is connected to the first slave server 110 as an additional visitor, and the third slave server is the first slave server 110 and is connected to the second slave server as an additional visitor.

After this, if a shutdown occurs in the master server 100, the connection of the new slave server is suspended until the master server 100 is restored, and the first slave server 110 to the third slave server that was previously connected are connected to the master server 110. Reconnection is attempted until the server 100 restores the connection. When the connection of the master server 100 is restored, the first slave server 110 to the third slave server transmit authentication packets in the order of connection. At this time, unlike in the past, the first slave server 110, the third slave server, and the second slave server can be connected to the master server 100 in order.

Then, the master server 100 sequentially transmits the fourth to sixth cluster numbers to the first slave server 110 to the third slave server according to the connected order. Through this process, the first slave server 110 received the 4th cluster number, the 3rd slave server received the 5th cluster number, and the 2nd slave server received the 6th cluster number. is Figure 5a.

Hereinafter, with reference to FIG. 3B, a restoration process when a second slave server is disconnected according to an embodiment of the present invention will be described in detail.

(S311) The master server 100 removes the second slave server from the slave server list after confirming the second slave server. (S312) The second slave server is the master server ( 100) and restores the connection. (S313) The master server 100 transmits a new cluster number to the second slave server. (S314) The second slave server reconnects to the first slave server 110. starts and restores the connection. (S315) The third slave server tries to reconnect to the second slave server and restores the connection. (S316)

A description of steps S311 to S316 will be given with reference to FIGS. 4C and 5B.

The description of FIG. 4c will be omitted since it has been mentioned in the description of steps S301 to S307.

If the second slave server is interrupted, the master server 100 checks the second slave server's interruption, removes the second slave server from the list of other slave servers, and then sends the second slave server to newly connected slave servers. Prepare to pass the excluded list. Afterwards, when the nth slave server (n is a natural number greater than 3) accesses, the second slave server delivers a list of excluded slave servers.

When a disconnection occurs, the second slave server tries to reconnect with the master server 100 to restore the connection, and the master server 100 transfers the fourth cluster number, which is a new cluster number, to the second slave server. When the disconnection of the connection to the first slave server 110 is confirmed, the second slave server restores the connection by attempting reconnection in the role of a visitor.

As for the third slave server, it waits for reconnection as a connection acceptor, and when the disconnection of the second slave server is confirmed, the third slave server attempts reconnection as a visitor and restores the connection. At this time, no role conflict occurs between the accessor and the access acceptor, so a separate role change does not occur. 5B shows the above process. In FIG. 5B, it can be confirmed that the cluster number of the first slave server 110 and the third slave server has not changed, but the cluster number of the second slave server has changed.

The aforementioned connection between the master server 100 and the second slave server means that the connection is temporarily disconnected due to a network problem. However, in addition to this, there are additional cases of clustering exception processing due to network disconnection that may occur in the slave server, which will be described in detail with reference to FIGS. 5C and 5D below.

FIG. 5C is an exemplary view when the master server 100 or the second slave server according to an embodiment of the present invention is terminated due to network disconnection.

When the clustering server is running with the configuration as shown in FIG. 4c, when the network connection of the second slave server is disconnected and the operation is terminated, the master server 100 lists the slave servers when the disconnection of the second slave server is confirmed. The second slave server is removed from , and the list from which the second slave server is excluded is transmitted to all slave servers after the nth slave server (n is a natural number greater than 3) that is newly connected.

Since the first slave server 110 performed the role of an access acceptor for the second slave server, when disconnection of the second slave server is confirmed, all information on the second slave server is removed.

Since the 3rd slave server performed the role of a visitor to the 2nd slave server, it tries to reconnect to the 2nd slave server when disconnection of the 2nd slave server is confirmed. At this time, if reconnection to the second slave server fails, the list of slave servers to be updated is requested from the master server 100 . In this process, reconnection is performed by dividing into a case where only the connection between the 2nd slave server and the 3rd slave server of the 3rd slave server is disconnected and a case where the 2nd slave server is disconnected from the master server 100 and all slave servers. .

If only the connection between the second slave server and the third slave server is disconnected, the third slave server requests the master server 100 to confirm the connection of the second slave server when reconnection to the second slave server fails, and the master server ( 100) informs the third slave server that the connection of the second slave server is being maintained. After confirming that the connection to the master server 100 is maintained by the second slave server, the third slave server repeatedly tries to reconnect to the second slave server and request connection confirmation to the master server 100 .

If the second slave server is disconnected from the master server 100 and all slave servers, and the third slave server fails to reconnect to the second slave server, it requests the master server 100 to confirm the connection of the second slave server, and , the master server 100 notifies the third slave server that the second slave server is disconnected. After confirming that the second slave server has disconnected from the master server 100, the third slave server suspends the reconnection and removes all information about the second slave server.

5D is an exemplary view when the master server 100 or the second slave server according to an embodiment of the present invention is restarted due to network disconnection.

In the description of FIG. 5D, the same content as the description of FIG. 5C above will be omitted.

When the clustering server is running with the configuration as shown in FIG. 4c, and the network connection of the second slave server is disconnected and restarted, the master server 100 and the first slave server 110, as mentioned in the previous description of FIG. 5c, Follow the same process as for

Since the 3rd slave server performed the role of a visitor to the 2nd slave server, it tries to reconnect to the 2nd slave server when disconnection of the 2nd slave server is confirmed. At this time, the reconnection process can be described by dividing into two cases: a case in which restarting and clustering of the second slave server are completed, and a case in which restarting or clustering of the second slave server is not completed.

When the restarting and clustering of the second slave server are completed, the second slave server receives a new cluster number, the fourth cluster number, from the master server 100 and a list of other slave servers, and transfers them to other slave servers in the list. Attempts to connect in the role of accessor. In this case, in the existing connection, the second slave server played the role of access acceptor for the third slave server, but after the restart, the role changes to the accessor. At this time, since the third slave server can also maintain the role of an accessor with respect to the second slave server, a role conflict occurs with the same accessor.

So, the 2nd slave server and the 3rd slave server compare each other's cluster numbers, and according to the priority, the 2nd slave server maintains the role of an accessor, and the 3rd slave server changes its role to an access acceptor and connects as an accessor role. Disconnect the network connection you had.

If the third slave server confirms that the connection with the second slave server is disconnected from the master server 100 before restarting or clustering of the second slave server is completed, the third slave server attempts reconnection with the second slave server. is stopped, and all information of the second slave server is removed. After that, when the second slave server is restarted, it receives a fourth cluster number, which is a new cluster number, and a list of other slave servers from the master server 100 . Then, an attempt is made to access the third slave server in the role of an accessor, and the third slave server completes connection and registration in the role of an access acceptor.

Hereinafter, referring to FIG. 6, a process in which a client device according to an embodiment of the present invention transmits a packet through a host device will be described in detail.

First, a step of defining the connection relationship between the host devices with the plurality of slave servers and the master server 100 is performed (S400). Since this process has been described with reference to FIG. 2 above, it will be omitted. Here, the host devices mean a plurality of servers, and is a concept including both slave servers and master servers.

Then, a plurality of client devices are registered with the host devices (S410). When the host devices are connected in step S400, the client device accesses one of the host devices and transmits an authentication packet. After receiving the authentication packet of the client device, the host device registers the client device as one of the client devices it manages. At this time, the devices managed by the host device are classified into two types: clustered host devices and client devices managed by itself.

7 is a diagram illustrating that a plurality of client devices are registered and connected to host devices. 7 shows a plurality of client devices (101, 102, ? ?,132,133) are connected.

The first client device 101, the second client device 102, and the third client device 103 are connected to the master server 100, and the 1-1 client device 111 is connected to the first slave server 110. , the 1-2nd client device 112 and the 1-3rd client device 113 are connected. The 2-1 client device 121, the 2-2 client device 122, and the 2-3 client device 123 are connected to the second slave server, and the 3-1 client device ( 131), the 3-2nd client device 132 and the 3-3rd client device 133 are connected.

Then, the client device transmits role information and packet type information to the host device (S420).

The client device transmits client role information and network packet type information to the connected host device, and the client role information is information defining which role the corresponding client is between a packet sender and a forwarder. At this time, the role information and the network packet type information are transmitted once, and are not additionally transmitted until the role information or the network packet type information is changed. For example, in order for a packet for processing group a to process group b, role information or network packet type information of a client device must be changed to a packet for processing group b. The host device operator operates the host device independently, and each user who wants to use the cloud service will want to use the corresponding host device for his/her own purpose. For example, the service is provided to the client devices of users who want to provide educational videos so that they can use the corresponding host device, and after the event to provide educational videos ends, the client devices of users who want to provide O2O platform services. It is also possible to provide a service so that users can use the corresponding host device. In this case, role information or network packet type information of each client device is changed, and additional transmission is required only when an event such a change occurs, and retransmission is not performed after transmission once.

In the prior art, role information and network packet type information are transmitted at regular intervals, but when they are transmitted at regular intervals, the load increases according to the number of transmissions. If it is possible to check whether the client is maintaining the connection, role information and network packet type information are transmitted only once and do not need to be transmitted repeatedly. However, when a new client device accesses the clustering, it transfers client role information and network packet type information managed by the corresponding host device.

In the present invention, each host device includes a main thread 210 (main thread) and a network packet processing thread 240 (network packet processing thread).

The main thread 210 manages host devices or client devices, registers and deregisters devices accessing it, and controls communication between host devices or packet flow between host devices and client devices. That is, flow control such as registration and deregistration of target devices regarding which host device or which client device a specific network payload packet should be delivered to is performed.

Network packet processing thread 240, also referred to as network payload packet processing thread, includes a transmit thread 222 and a receive thread 221. A transmitting thread 222 and a receiving thread 221 are each assigned to one host device. This is to ensure that a transmission problem or a reception problem between network communications occurring in a specific device does not affect the communication of other devices. A packet received from a specific host device is processed by the receiving thread 221, and among the packets, a control packet is transferred to the main thread 210, and the page is sent to the sending thread 222 of the target device designated by the main thread 210. Forward the load packet. A description thereof will be made with reference to FIG. 8 .

8 is an exemplary diagram for network packet control through thread separation in an nth slave server according to an embodiment of the present invention. In the following description, the main thread 210 shown in FIG. 8 is set as the n-th slave server including the network packet processing thread 220 for the first slave server and the network packet processing thread 230 for the second slave server. to explain. That is, each slave server includes a network packet processing thread for other slave servers.

When a packet including a control packet and a payload packet is transferred from the first slave server to the nth slave server, the packet is transferred to the receiving thread 221 of the network packet processing thread 220 for the first slave server. The receiving thread 221 transfers a control packet among the transferred packets to the main thread 210 . The main thread 210 checks which host device the payload packet is going to through the control packet, and recognizes that the corresponding target host device is the second slave server. Next, the network packet processing thread 220 for the first slave server transfers the payload packet received through the reception thread 221 to the transmission thread 222 of the network packet processing thread 230 for the second slave server. do. The transmitting thread 222 of the network packet processing thread 230 for the second slave server that has received the payload packet transfers the payload packet to the second slave server.

Packet movement between slave servers described in FIG. 8 requires guaranteeing packet integrity and minimizing delay during packet transmission. In order to guarantee packet integrity, the entire packet is divided into MTU-unit frames and transmitted to the network, which will be described with reference to FIG. 9 below.

9 is an exemplary diagram illustrating a configuration of an MTU according to an embodiment of the present invention. When MTU transmits a packet, it divides it into a maximum size of 1500 bytes and sends it. Packet configuration includes frame number, start of packet, end of packet, frame size, and information about the frame. In FIG. 9, a frame of 4000 bytes is divided into 1500 bytes, 1500 bytes, and 1000 bytes, respectively, and frame numbers are divided into 0, 1, and 2.

The host device shares client role information and packet type information with other host devices (S430).

A host device manages client role information and network packet type information received from client devices accessing it, and transmits all client role and network packet type information managed by itself to other clustered host devices.

The host device shares the role information received from other host devices with the client devices it manages (S440).

The host device transmits the role information of client devices among the role information of client devices and network packet type information received from other host devices to all client devices managed by the host device. If client role information is delivered to client devices, all clients connected to the clustering service can directly manage the desired remote client device.

Since network packet type information can deliver specific network packets to specific client devices only through communication between host devices, the network packet type information is managed only by the host device itself without being transmitted to the client devices. That is, only the host device needs to know where this packet is going.

The role information of the client device transmitted from the host devices can manage client devices having a role desired by a specific client device.

For example, the role of the 2-1st client device 121 registered in the 2nd slave server is registered and managed by the 1-1st client device 111 registered in the 1st slave server 110. When the driving of the client device 121 is terminated, the second slave server notifies the first slave server 110 that the driving of the 2-1 client device 121 is terminated, and the 1-1 client device 111 When the operation of the 2-1 client device 121 is terminated, the management information is removed, and the 2-1 client device 121 searches for and communicates with other client devices capable of performing the same role as the ongoing service.

The client device forwards the specific packet to the host device along with packet type information (S450). The host device refers to the pre-stored packet type information for each client device and sends it to another host device connected to the client device that can process the specific packet. Deliver. (S460)

For example, the 2-1 client device 121 transfers a specific type of packet to the second slave server along with network packet type information. Then, the second slave server transfers client devices capable of receiving and processing packets of the corresponding type to the registered host devices, and the received host devices find and deliver client devices capable of receiving and processing packets of the corresponding type.

When a client device or a host device receives a packet, it can identify a control packet and a payload packet by analyzing the front structure of the packet, and through this, which thread among the main thread 210 and the network packet processing thread 240 decide whether to proceed through Classification of the types of control packets and payload packets will be described with reference to FIG. 10 below.

10 is an exemplary diagram illustrating classification of packet types according to an embodiment of the present invention.

When each packet data is divided into a control packet and a payload packet, the delay between packet transmissions is minimized when the division proceeds quickly. In the present invention, information for classifying packet types is placed at the front of all packets, and control packets are classified as "0" and payload packets as "1". Through this, when a client device or a host device receives a packet, it can minimize a delay that may occur between packet analysis by distinguishing a control packet or a payload packet with only some data at the beginning of the packet without analyzing other information.

An embodiment of the present invention may be implemented in the form of a recording medium including instructions executable by a computer, such as program modules executed by a computer. Computer readable media can be any available media that can be accessed by a computer and includes both volatile and nonvolatile media, removable and non-removable media. Also, computer readable media may include all computer storage media. Computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data.

An embodiment of the present invention may be implemented in the form of a recording medium including instructions executable by a computer, such as program modules executed by a computer. Computer readable media can be any available media that can be accessed by a computer and includes both volatile and nonvolatile media, removable and non-removable media. Also, computer readable media may include all computer storage media. Computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data.

Although the methods and systems of the present invention have been described with reference to specific embodiments, some or all of their components or operations may be implemented using a computer system having a general-purpose hardware architecture.

The above description of the present invention is for illustrative purposes, and those skilled in the art can understand that it can be easily modified into other specific forms without changing the technical spirit or essential features of the present invention. will be. Therefore, the embodiments described above should be understood as illustrative in all respects and not limiting. For example, each component described as a single type may be implemented in a distributed manner, and similarly, components described as distributed may be implemented in a combined form.

The scope of the present invention is indicated by the following claims rather than the detailed description above, and all changes or modifications derived from the meaning and scope of the claims and equivalent concepts should be construed as being included in the scope of the present invention. do.

100: master server 101: first client device
102: second client device 103: third client device
110: 1st slave server 111: 1-1st client device
112: 1st-2nd client device 113: 1st-3rd client device
120: second slave server 121: 2-1 client device
122: 2-2nd client device 123: 2-3rd client device
130: 3rd slave server 131: 3-1st client device
132: 3-2nd client device 133: 3-3rd client device
210: main thread 220: network packet processing thread for the first slave server
221: receiving thread 230: network packet processing thread for the second slave server
222: sending thread 240: network packet processing thread

Claims (14)

In the large-capacity network packet management method for client device registration and ultra-low delay,
(a) registering a plurality of client devices for each of the host devices;
(b) the client device transmitting client role information and network packet type information to be processed to a host device, and the host device sharing the client role information and network packet type information with other host devices;
(c) sharing, by the host device, client role information among information received from other host devices with a client device managed by the host device; and
(d) When the client device forwards a specific packet to the host device along with network packet type information, the host device may process the specific packet by referring to pre-stored network packet type information for each client device. transferring the specific packet to another host device connected to the other client device so that the specific packet can be processed;
Including, wherein the host devices are composed of a plurality of slave servers and master servers, and the steps (b) to (d) are performed for each host device,
A large network packet management method for client device registration and ultra-low latency. According to claim 1,
The devices managed by the host device correspond to any one of the clustered host devices and the client devices that delivered the authentication packet to the host device.
A large network packet management method for client device registration and ultra-low latency. According to claim 1,
In step (b), the client device additionally transmits client role information and network packet type information when the client role information and network packet type information to be processed are changed.
A large network packet management method for client device registration and ultra-low latency. According to claim 1,
Before step (a),
(a-1) the first slave server and the second slave server transmitting access and authentication packets to the master server;
(a-2) transmitting, by the master server, a cluster number and a list of other accessible slave servers to the first slave server and the second slave server; and
(a-3) When the first slave server accesses the master server earlier than the second slave server, the second slave server transmits an access and authentication packet to the first slave server, and the first slave server the server accepting the connection;
Including more,
The cluster number is sequentially issued according to the order in which the master server transmits access and authentication packets of the slave server,
In the connection between the slave servers, the slave server at the point in time at which the slave server connects to the master server and transfers the authentication packet takes the role of an access acceptor, and the slave server connects to the master server and authenticates. The slave server at which the packet delivery time is late becomes the accessor,
A large network packet management method for client device registration and ultra-low latency. According to claim 1,
In step (a),
transmitting an authentication packet after a client device accesses any one of host devices composed of a plurality of clustered slave servers and master servers; and
registering, by the host device receiving the authentication packet, the client device as one of client devices managed by the host device;
Which includes,
A large network packet management method for client device registration and ultra-low latency. According to claim 1,
The network packet type information is
Shared between the host devices, but not shared among the client devices;
The host device transmits the client role information when a new client device accesses the clustering server.
A large network packet management method for client device registration and ultra-low latency. According to claim 1,
After step (d),
storing, by the 1-1st client device registered in the 1st slave server, the role of the 2-1st client device registered in the 2nd slave server;
notifying, by the second slave server, the end of driving of the 2-1 client device to the first slave server when the driving of the 2-1 client device is terminated; and
The 1-1 client device removes management information from the list of stored client devices according to the end of driving of the 2-1 client device, and selects another client device capable of the same role by referring to pre-stored client role information. Step of finding and requesting communication;
Which includes,
A large network packet management method for client device registration and ultra-low latency. According to claim 1,
Each host device and each client device:
main thread; and
a network packet processing thread including a receiving thread and a sending thread;
Which includes,
A large network packet management method for client device registration and ultra-low latency. According to claim 8,
A payload packet is a packet generated by a client device and transmitted to be processed by another client device.
The control packet is a packet specifying a path of the payload packet,
Each host device and each client device,
When a control packet is received, it is processed through the main thread,
When a payload packet is received, it is processed through a network packet processing thread.
A large network packet management method for client device registration and ultra-low latency. According to claim 9,
Each server includes a network packet processing thread for each of the other servers.
When a server includes a network packet processing thread (including a sending thread and a receiving thread) for the first slave server and a network packet processing thread (including a sending thread and a receiving thread) for the second slave server,
When any one of the servers receives a packet from the first slave server, the payload packet is received through a receiving thread for the first slave server,
When the control packet indicates transmission to the second slave server,
transferring the payload packet to the second slave server through a transmission thread for the second slave server;
A large network packet management method for client device registration and ultra-low latency. According to claim 9,
The client device or host device, for packet integrity
The entire packet is divided into frames in units of MTU and transmitted to the network,
The packet includes a number of frames, a start of a packet, an end of a packet, and a size of a frame,
The MTU size does not exceed 1,500 bytes,
A large network packet management method for client device registration and ultra-low latency. According to claim 11,
Information containing identification information on whether the corresponding packet is a control packet or a payload packet is placed at the front of the corresponding packet,
A large network packet management method for client device registration and ultra-low latency. According to claim 9,
The client device or the host device receives the packet, identifies whether it is a control packet or payload packet by analyzing only the first structure of the packet, and determines which thread to process through the main thread and the network packet processing thread. sign,
A large network packet management method for client device registration and ultra-low latency. In a large-capacity network packet management system for client device registration and ultra-low latency,
a plurality of host devices and a plurality of client devices;
The apparatus executes a method of performing bulk network packet management;
The above method
(a) registering a plurality of client devices for each of the host devices;
(b) the client device transmitting client role information and network packet type information to be processed to a host device, and the host device sharing the client role information and network packet type information with other host devices;
(c) sharing, by the host device, client role information among information received from other host devices with a client device managed by the host device; and
(d) When the client device forwards a specific packet to the host device along with network packet type information, the host device may process the specific packet by referring to pre-stored network packet type information for each client device. transferring the specific packet to another host device connected to the other client device so that the specific packet can be processed;
Including, wherein the host devices are composed of a plurality of slave servers and master servers, and the steps (b) to (d) are performed for each host device,
High-capacity network packet management system for client device registration and ultra-low latency.

Images