KR101397656B1 - Method for Managing Duplication of Server Using L4 Switch - Google Patents

Abstract

The present invention relates to a server duplication management method using an L4 switch.

In the present invention, the master and backup setting states of the L4 switch and the active and standby states of the L4 switch in the redundant server structure are grasped and the master, the backup setting state, the active state, and the standby state of the two servers, Provides a table structure and management method that can be grasped.

The present invention can expect the effect of extracting pair equipment of L4 switch, L4, master of server, backup setting state, and active standby switching state by using L4 switch.

L4 switch, server redundancy management, active, standby, master, backup

Description

Method for managing server redundancy using L4 switch [

FIG. 1 is a view illustrating a configuration of a server duplication management system using an L4 switch according to an embodiment of the present invention.

2 is a diagram for explaining the redundancy configuration (physical configuration 2a and logical configuration 2b) of the L4 switch according to the embodiment of the present invention.

3 is a diagram for explaining a method of configuring a backup server in the SLB of the L4 switch according to the embodiment of the present invention.

FIG. 4 is a view for explaining a structure of an active standby structure according to an embodiment of the present invention and a VRRP and an SLB table for state grasp.

5 is a diagram for explaining a state management method of an L4 switch according to an embodiment of the present invention.

6 is a diagram for explaining a state management method of a server using an L4 switch according to an embodiment of the present invention.

The present invention relates to a server redundancy management method using an L4 switch, and more particularly, to a server redundancy management method and an active standby redundancy management method using an L4 switch.

In the network (router, switch, etc.) facility, the demand of the user is diversified according to the spread of the Internet, and the introduction of the server becomes a necessity when providing the function that can not be provided to the customer.

Since the configuration of the system on the server can be easily changed during operation, it is necessary to introduce a redundancy structure in order to minimize the downtime and to provide uninterrupted service. The system redundancy structure is active-active and active-standby.

In active-active mode, two servers are responsible for 50/50 of total service processing requests. This method is advantageous in that all the devices can be utilized without idle servers because load is distributed to two servers. On the other hand, if one server fails, the other one server must perform all processing, There is a problem.

The active-standby method is a method in which one server processes all service requests and, if a failure occurs, the other apparatuses that are waiting are processing service requests.

This method can be performed without interruption in case of failover at a time that the customer can not feel and can tolerate.

Such an active-standby scheme is preferred because of its uniform serviceability, simplicity in system design and operation.

Systems that are closely related to customers have adopted the active-standby method, which is expensive in terms of equipment introduction but superior in terms of service. However, there is a problem in that the operator can not easily recognize the active standby switching occurring in the server operation, thereby failing to appropriately perform the redundancy management.

In order to solve such a problem, the present invention provides a server redundancy management method using an L4 switch.

According to an aspect of the present invention, there is provided a server redundancy management method comprising the steps of: (a) searching a real server table including a real server index for a first real server indicating the real server index; (b) retrieving a group index matched with the searched real server index through a real server group table connecting a real server layer and a group layer; (c) retrieving a backup group index matching the searched group index through a group table indicating the group hierarchy; And (d) a second real server indicating a real server index-backup server corresponding to the retrieved backup group index using the real server group table and the real server table, And searching for a pair server of the second real server.

A server duplication management method according to an aspect of the present invention includes: (a) searching a real server index in a real server table indicating a real server hierarchy; (b) searching a group index indicating a group hierarchy and a group index matching the searched real server index using a real server group table connecting the group hierarchy and the real server hierarchy; And (c) searching a virtual server index matching the retrieved group index by using a virtual server table representing a virtual server layer, a virtual service table connecting the group table and the virtual server table, And determining the setting state of the searched real server index as a master or a backup by searching the corresponding virtual IP.

A server duplication management method according to an aspect of the present invention includes the steps of: (a) searching a virtual server index in a virtual server table indicating a virtual server layer; (b) searching a group index indicating a group hierarchy and a group index matched with the searched virtual server index using a virtual service table connecting the virtual server table and the group table; (c) searching a real server index indicating a real server hierarchy and a real server index matched with the retrieved group index using a real server group table connecting the group hierarchy and the real server hierarchy; And (d) determining a current status of the searched real server index as active or standby by searching a virtual service status matched with the searched virtual server index using a virtual server port table indicating a service port status of the virtual server, .

A server duplication management method according to an aspect of the present invention includes the steps of (a) receiving a first L4 switch including a first virtual router redundancy protocol table and a second virtual router redundancy protocol table including a second virtual router redundancy protocol table, Transmitting and receiving an advertisement packet between L4 switches; (b) searching for a pair IP by searching for an interface IP having the same virtual IP using the first virtual router redundancy protocol table and the second virtual router redundancy protocol table; (c) comparing the priority of the first virtual router redundancy protocol table with the priority of the second virtual router redundancy protocol table, and comparing the setting state of the first L4 switch and the priority of the second L4 switch Determining a setting state of the master device or the backup device; And (d) searching for a state value of the first virtual router redundancy protocol table and a state value of the second virtual router redundancy protocol table, and comparing the current state of the first L4 switch and the current state of the second L4 switch Judging the status as active or backup.

Hereinafter, 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. The present invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. In order to clearly illustrate the present invention, parts not related to the description are omitted, and similar parts are denoted by like reference characters throughout the specification.

Throughout the specification, when an element is referred to as "comprising ", it means that it can include other elements as well, without excluding other elements unless specifically stated otherwise. Also, the terms " part, "" module," " module, "and " block" refer to units that process at least one function or operation, Lt; / RTI >

Now, a server redundancy management method using an L4 switch according to an embodiment of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a view illustrating a configuration of a server duplication management system using an L4 switch according to an embodiment of the present invention.

The server duplication management system according to an embodiment of the present invention includes a Giga Switch Router (GSR) 100, first and second L4 switches 200 and 210, and first and second servers 300 .

The first and second servers 300 having the same function are grouped into a service and processed. If there is a server group processing another service, it is connected to the L4 switches 200 and 210. [

The equipment configuration follows a hierarchical structure and different routes are applied to each section.

The L4 switches (200, 210) and the Giga switch routers are static routing sections, and both uplink and downlink traffic are configured in a floating static manner in preparation for failure.

Therefore, if one routing path fails, the traffic is communicated to the other path.

The connection links between the L4 switches 200 and 210 and the Giga switch routers are in different segments (Broadcasting Domains). The lower connection link for performing redundancy to the server and the interlink between the L4 switches (200 and 210) are all in the same segment.

The default router of each server for uplink traffic is designated as a virtual IP of a Virtual Router Redundancy Protocol (VRRP) operating between the L4 switches 200 and 210.

Down traffic of the L4 switches 200 and 210 is sent to an Active Real Server belonging to the virtual IP of the server load balancing (SLB) set in L4.

FIG. 2 is a diagram for explaining the redundancy configuration (physical configuration 2a and logical configuration 2b) of the L4 switches 200 and 210 according to the embodiment of the present invention.

VRRP is a protocol defined in RFC3768 and proposed to overcome the disadvantage of the static default router set in the server.

That is, in case of a server, in order to solve the problem of network disconnection when a default router set in the server fails, an IP set as a default router is virtually given, so that the IP is always recognized as a live interface IP.

Ownership of the virtual IP is provided by two paired L4 switches (200, 210). Each L4 switch 200 and 210 receives the same VRRP packet with the same virtual router ID.

The interval in which the VRRP protocol operates is an Ethernet interval that is a broadcasting domain. In addition, the port and L4 interlink to the server are set to the same Vlan.

The virtual router of each L4 determines the master according to the priority, that is, the priority value. If the backup device does not periodically receive an advertisement packet from the master device, the backup device determines that the counterpart device is dead and switches itself to the active state.

Each interface of the paired L4 device sets the virtual ID and the virtual IP equal to each other. One interface can have multiple virtual IPs for SLB purposes as well as virtual IPs for server-side default routers.

Next, referring to FIG. 3, an SLB configuration method of the L4 switches 200 and 210 will be described in detail.

3 is a diagram for explaining a method of configuring a backup server in the SLB of the L4 switches 200 and 210 according to the embodiment of the present invention. 3a is a method of directly designating a backup server for a real server according to an embodiment of the present invention, and 3b is a method of designating a backup group for a group according to an embodiment of the present invention.

SLB supports load balancing and failure failover by associating two or more Real Servers with one virtual IP.

Each virtual IP and service port pair corresponds to a group one by one. A group is a basic unit that shares a health check and a metric for a server. The real servers belonging to the group are connected according to the load balancing method such as Roundrobin, Hash, LeastConnection, etc. as specified in the metric.

As shown in FIG. 3A, the connection between the virtual IP and the real server for server load balancing is performed through group designation in the middle.

When a group is specified for a virtual IP and a service port, a service request is transmitted to a real server belonging to the group. Groups related to the service are configured differently for each service, and one server belonging to the group is designated by the same server.

The server designated as the backup handles the connection request of the client when the service port failure occurs in the server belonging to the group.

The method of directly designating the backup server for such a real server has the characteristic that even if a failure occurs in one service port, the entire service is not transferred to another server but only the failed service is switched to the backup server .

3b, a virtual server for each service is mapped to one group, and even if one of the service ports of the server belonging to the group, the entire service request is transferred to the server designated in the backup backup group do.

A method of specifying a backup group for such a group is used when a plurality of service ports belonging to a group are related to each other. If a failure occurs in one service port, a block is blocked even if the remaining service ports are normal, To another server.

Next, the structure of the VRRP and SLB tables 400 and 500 will be described in detail with reference to FIG.

FIG. 4 is a view for explaining a structure of an active standby structure and VRRP and SLB tables 400 and 500 for status grasp according to an embodiment of the present invention.

The VRRP table 400 includes an Equip Identifier 410, a virtual IP 420, an interface IP 430, a Priority 440, and a Status 450.

The equipment ID 410 is a unique ID that can identify the equipment name or equipment.

The virtual IP 420 is a virtual IP address of the virtual router, and is a virtual IP address 420 to be exchanged through the VRRP protocol in the virtual router.

The interface IP 430 is an interface IP 430 to which the VRRP protocol is transmitted, and is an IP of a physical interface attached to a segment in which the VRRP protocol operates.

The priority 440 has a value ranging from 0 to 255. As the value increases, the priority 440 has priority. Based on this value, the master 440, which is the setting state of the L4 switches 200 and 210, determines whether or not to backup.

The state 450 indicates the state value of the virtual router, indicating whether the current state is active or standby. Based on this value, the active state, which is the current state of the L4 switches 200 and 210, is determined.

The SLB table 500 includes a server configuration table 510, a connection configuration table 550, and a current status table 580.

The server configuration table 510 includes a real server table, a group table, and a virtual server table.

The real server table 520 includes a real server index 522, a real IP 524, and a backup server index 526 as components.

The group table 530 includes a group index 532, a backup group index 534, a metric 536 for load distribution, and a health check 538 for detecting failure Element.

The virtual server table 540 includes a virtual server index 542 and a virtual IP 544 as components.

The connection configuration table 550 includes a real server group table 560 and a virtual service table.

The real server group table 560 includes a real server index 562 and a group index 564 as components. The virtual service table 570 includes a virtual server index 572, a service port 574, and a group index 576 as components.

The mapping between the real server and the group is performed through the real server group table 560. The mapping between the group and the virtual server is performed through the virtual service table 570.

Each service (virtual IP 544, service port 574) of the virtual service is mapped to a group.

The status of the virtual service has Running / Blocked / Failed according to the status of the service port 574 up / down.

The current state table 580 includes a virtual server port table 590 representing a virtual server index 592, a virtual service index 594, a real server index 596, a service port 598 and a virtual service state 599 .

Here, the real server index 522 of the real server table 520, the real server index 562 of the real server group table 560, and the real server index 596 of the virtual server port table 590 are stored in the same real server IP.

The group index 532 of the group table 530, the group index 564 of the real server group table 560 and the group index 576 of the virtual service table 570 mean the same group.

The virtual server index 542 of the virtual server table 540, the virtual server index 572 of the virtual service table 570 and the virtual server index 592 of the virtual server port table 590 are stored in the same virtual server IP address .

The service port 574 of the virtual service table 570 and the service port 598 of the virtual server port table 590 refer to the same port.

Next, referring to FIG. 5, a method of managing states of the L4 switches 200 and 210 among the server redundancy management methods using the L4 switches 200 and 210 will be described in detail.

5 is a diagram for explaining a state management method of the L4 switches 200 and 210 according to the embodiment of the present invention.

5, the first L4 switch 200 and the second L4 switch 210 respectively include the VRRP table 400 and the SLB table 500 described in FIG.

The first L4 switch 200 and the second L4 switch 210 include a first control unit and a second control unit, respectively, and the functions of the first control unit and the second control unit are the same. The first control unit and the second control unit transmit and receive a VRRP advertisement packet (S100), respectively. The VRRP table 400 and the SLB table 500 are used to extract pairs of L4 and real servers, And manages the setting state, the active state in the operation, and the standby state.

Therefore, for convenience of description, the first control unit and the second control unit will be referred to as a control unit hereinafter.

The control unit searches for a pair IP for each interface IP 430 in which the VRRP protocol operates among the L4 interfaces. In other words, the control unit recognizes the pair IP by searching the interface IP 430 having the same VRRP virtual IP 420 in the VRRP table 400 of each of the first L4 switch 200 and the second L4 switch 210 S102).

The control unit compares the VRRP table 400 of the first L4 switch 200 with the priority 440 in the VRRP table 400 of the second L4 switch 210 and has a high value interface The L4 switch is determined to be the master device and the remaining L4 switch is determined to be the backup device (S104).

The controller determines the current status (Active, Standby) using the status values of the two interfaces sharing the same VRRP virtual IP 420 (S106).

Next, referring to FIG. 6, a method of managing the state of the server among the redundancy management methods using the L4 switches 200 and 210 will be described in detail.

6 is a diagram for explaining a state management method of a server using L4 switches 200 and 210 according to an embodiment of the present invention.

3, 4 and 6, a description will be given of a state management method of a server using L4 switches 200 and 210 for detecting pairs of equipments, a setting state and a current state of a real server.

As shown in FIG. 3, when a backup server for a real server is directly designated and a connection between the first real server and the second real server is performed through group designation in the middle, Should be considered.

(1) When a backup server for a real server is directly specified, a method for searching the setting status and the current status of a pair server and a real server is as follows.

1) How to search for a pair server is as follows.

The control unit refers to the real server table 520 in the server configuration table 510 of the SLB table 500 to determine that the backup server index 526 is 0 for the specified real server index 522 . If the backup server index 526 is not 0, the controller determines that a backup server exists and refers to the real server index 522 and the backup server index 526 to determine whether there is a backup server, (S200).

2) How to search the setting status of real server is as follows.

The control unit determines whether the virtual service of each real server is a master or a backup. That is, the controller moves to the upper layer in the order of the real server layer, the group layer, and the virtual server layer to check whether the virtual IP 544 exists. If there is no host group or the virtual IP 544 does not exist, (S202). If the virtual IP 544 exists, the setting state of the real server is determined as the master (S204).

More specifically, the controller searches the real server table 520 for a real server index 522 (S202), and determines whether a backup server index 526 exists.

If the backup server index 526 exists, the control unit determines that the setting of the real server index 522 is a backup because the parent group does not exist (S204).

If the backup server index 526 does not exist, the controller refers to the real server group table 560 and the group table 530 based on the retrieved real server index 522 and searches the group index 532 ). Then, the control unit searches the virtual server index 572 by referring to the virtual service table 570 based on the retrieved group index 532. The controller looks up the virtual server table 540 based on the searched virtual server index 572 and finds a virtual IP 544 corresponding to the virtual server index 572 in step S208. The controller determines the setting state of the searched real server index 522 as a master.

3) How to retrieve the current state of the real server is as follows.

The control unit joins the real server table 520, the real server group table 560, the group table 530, the virtual service table 570 and the virtual server table 540, The real server corresponding to the real server 574 is searched. The control unit then retrieves the virtual service status 599 based on the virtual IP 544 and the service port 574 in the virtual server port table 590.

If the virtual service state 599 is Running, the controller determines that the current state of the detected real server is active. If the virtual service state 599 is Blocked or does not have a value, the controller determines the current state of the server as a standby state.

In more detail, the control unit searches the virtual server table 540 for a virtual server index 542 including the virtual IP 544 (S210). The control unit searches the service port 574 and the group index 532 corresponding to the virtual server index 542 with reference to the virtual service table 570 and the group table 530 based on the retrieved virtual server index 542 (S212).

The controller refers to the real server group table 560 based on the retrieved group index 532 and searches the real server index 562 corresponding to the retrieved group index 532 in operation S214. The control unit refers to the real server table 520 based on the searched real server index 562 and searches the real IP 524 and the backup server index 526 corresponding to the searched real server index 562 at step S216.

The control unit then retrieves the virtual service status 599 by referring to the virtual server port table 590 based on the virtual IP 544, the service port 574, and the real server index 522.

The controller determines that the current state of the real server is Running / Blocked / Failed according to the virtual service state 599. Here, if the current state of the real server is running, the current state of the backup server becomes a block or a fail.

(2) When the connection between the first real server and the second real server is performed through group designation in the middle, a method of searching the setting state and the current state of the pair server and the real server is as follows.

1) How to search for a pair server is as follows.

The control unit refers to the real server group table 560 based on the real server index 522 given in the real server table 520 among the server configuration table 510 and searches the group index 564 having the real server as a member (S300, S302). Then, the control unit searches the backup group index 534 by referring to the group table 530 based on the retrieved group index (representing the first real server) 564 (S304).

When searching the backup group index 534 in the group table 530, the controller refers to the real server group table 560 based on the group index 532 including the retrieved backup group index 534, The real server index 562 corresponding to the real server index 532 is searched (S306).

The controller searches the real server table 520 for the real IP 524 corresponding to the searched real server index (representing the second real server) 522. Accordingly, the control unit finds a pair server in which connection between the virtual IP 544 and the real server is performed through group designation in the middle.

2) How to search the setting status of real server is as follows.

The control unit searches the real server index 522 of the real server table 520 at step S308 and searches the group index 564 by referring to the real server group table 560 based on the searched real server index 522 (S310).

When the group index 564 is found, the control unit determines whether a backup group index 534 exists in the group table 530. If the backup group index 534 exists, the controller determines that the setting of the retrieved real server index 522 is a backup because the virtual IP 544 does not exist at step S312.

The controller refers to the virtual service table 570 and the virtual server table 540 on the basis of the retrieved group index 564 in the case where the backup server index 526 does not exist, 544 (S314), and determines the setting state of the searched real server index 522 as a master.

The control unit controls the virtual IP 544 for the real server 544 and the service port 540 via the real server table 520, the real server group table 560, the group table 530, the virtual service table 570, (574) is finally searched, the setting state is determined as the master.

3) How to retrieve the current state of the real server is as follows.

The control unit searches the virtual server table 540 for a virtual server index 542 including the virtual IP 544 (S316). The controller searches the service port 574 and the group index 576 corresponding to the virtual server index 542 by referring to the virtual service table 570 based on the retrieved virtual server index 542 at step S318.

The control unit searches the backup group index 534 by referring to the group table 530 based on the retrieved group index 576. [ If the retrieved backup group index 534 does not exist, the controller refers to the real server group table 560 and searches the real server index 562 corresponding to the retrieved group index 576 at step S320.

The controller refers to the real server table 520 based on the retrieved real server index 562 and searches for the real IP 524 corresponding to the retrieved real server index 562.

If the retrieved backup group index 534 exists, the controller refers to the real server group table 560 and the real server table 520 based on the group index 532 indicating the backup group index 534, The real IP 524 corresponding to the server index 522 and the real server index 522 is searched (S322 and S324).

In this way, the control unit accesses the virtual service status 599 from the virtual server port table 590 of the current status table 580 using the virtual IP 544 and the service port 574 and the real server of the group real server / . The control unit becomes a standby state when the server in which the virtual service state 599 is running is active and is a block or has no value.

The embodiments of the present invention are not limited to the above-described apparatuses and / or methods, but may be implemented through a program for realizing functions corresponding to the configuration of the embodiment of the present invention, a recording medium on which the program is recorded And such an embodiment can be easily implemented by those skilled in the art from the description of the embodiments described above.

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, It belongs to the scope of right.

According to the above-described configuration, the present invention can be expected to have an effect of grasping the setting state and the current state of the server with respect to server switching occurring in the duplicated system.

The present invention can expect the effect of extracting pair equipment of L4 switch, L4, master of server, backup setting state, and active standby switching state by using L4 switch.

Claims (9)

A virtual router redundancy protocol table and a server load balancing table, wherein the server load balancing table includes a real server table, a real server group table, and a group table. A method for performing duplication management for a server by transmitting and receiving an advertisement packet to and from another L4 switch, (a) searching a first real server indicating the real server index in a real server table including a real server index; (b) retrieving a group index matched with the searched real server index through a real server group table connecting a real server layer and a group layer; (c) retrieving a backup group index matching the searched group index through a group table indicating the group hierarchy; And (d) a second real server indicating a real server index-backup server corresponding to the retrieved backup group index using the real server group table and the real server table, Searching for a pair server of the second real server Wherein the server is a server. Virtual Router Redundancy Protocol table and Server Load Balancing table - where the server load balancing table contains the real server table, real server group table, group table, virtual server table, and virtual service table. The method comprising the steps of: receiving, by the L4 switch, an advertisement packet transmitted from another L4 switch and performing redundancy management on the server, (a) searching a real server index in a real server table indicating a real server hierarchy; (b) searching a group index indicating a group hierarchy and a group index matching the searched real server index using a real server group table connecting the group hierarchy and the real server hierarchy; And (c) searching for a virtual server index matching the retrieved group index by using a virtual server table representing a virtual server layer, a virtual service table connecting the group table and the virtual server table, Determining a setting state of the searched real server index as a master or a backup Wherein the server is a server. 3. The method of claim 2, And if the backup server index exists in the real server table, determining that the set status of the retrieved real server index is a backup because there is no higher level group. 3. The method of claim 2, And if the backup group index exists in the group table, determining that the set status of the retrieved real server index is a backup because there is no virtual IP of the virtual server. The Virtual Router Redundancy Protocol table and the Server Load Balancing table, where the server load balancing table includes the real server table, the real server group table, the group table, the virtual server table, the virtual service table, A method for performing redundancy management for a server by transmitting and receiving an advertisement packet to and from another L4 switch, the L4 switch including a virtual server port table, (a) searching a virtual server index in a virtual server table representing a virtual server layer; (b) searching a group index indicating a group hierarchy and a group index matched with the searched virtual server index using a virtual service table connecting the virtual server table and the group table; (c) searching a real server index indicating a real server hierarchy and a real server index matched with the retrieved group index using a real server group table connecting the group hierarchy and the real server hierarchy; And (d) determining a current status of the retrieved real server index as active or standby by retrieving a virtual service status matched with the retrieved virtual server index using a virtual server port table indicating a service port status of the virtual server Wherein the server is a server. 6. The method of claim 5, In the step (b) Searching for a backup group index in the group table; And If the backup group index is searched, searching for a real server index-backup server corresponding to the searched backup group index using the real server group table and the real server table Wherein the server is a server. 6. The method of claim 5, In the step (c) Searching for a backup server index for the searched real server index through the real server table Wherein the server is a server. A method for performing redundancy management for a server, the first L4 switch including a Virtual Router Redundancy Protocol table, (a) transmitting and receiving an advertisement packet with a second L4 switch; (b) searching for a pair IP by searching for an interface IP having the same virtual IP using the virtual router redundancy protocol table of the first L4 switch and the virtual router redundancy protocol table of the second L4 switch; (c) comparing the priority of the virtual router redundancy protocol table of the first L4 switch with the priority of the virtual router redundancy protocol table of the second L4 switch, And determining the setting state of the second L4 switch as a master device or a backup device; And (d) a state value of the first virtual router redundancy protocol table of the first L4 switch and a state value of the virtual router redundancy protocol table of the second L4 switch are searched for and the current state of the first L4 switch Determining the current state of the second L4 switch as active or backup Wherein the server is a server. 9. The method of claim 8, Wherein the first L4 switch and the second L4 switch comprise a server load balancing table associated with server load balancing, The server load balancing table includes: A server configuration table including a real server table associated with the real server layer, a group table associated with the group layer, and a virtual server table associated with the virtual server layer; A connection configuration table for providing a mapping relationship between the real server table, the group table, and the virtual server table; And A current state table indicating the state of the virtual service Wherein the server is a server.

Images