KR20020022923A - Method for Database of a Load Balancing Equipment and Its System - Google Patents

Abstract

PURPOSE: A method and a system for configuring a database of load balancing are provided to realize a high performance and high operation by configuring a database for managing the connection and session information of a load balancing in a proper way. CONSTITUTION: The identifier-address conversion module(11) plays a role of turning over the inputted identifier to a new information allocating space and turning over the previously set address of storage. The identifier combines and classifies the Internet address of the user connection, the connection information and the user information. The identifier-address conversion module returns a defined address value according to the information such as a define method of identifier, a number of real server, a size of storage, a unit size of group management, and so on by being operated with a combination of such as a hashing algorithm. An identifier information connection space(12) is classified by the classifier, stores the information of each connection and connection group and manages the connection to an information allocation space(13). The information allocation space manages the real storage as a buffer pool in order to efficiently manage the storage space. An identifier information recovery block(14) plays a role of recovering the identifier information according to the operation mode of database and the identifier-address conversion module.

Description

Method for Database of a Load Balancing Equipment and Its System

A load balancer is often used to connect multiple servers when assigning them to a single virtual server by connecting them to each other through a clustering method and gaining high performance and reliability.

The present invention corresponds to the technical field related to a method and system for constructing a database for managing session information or connection information inside a load balancer as shown in FIG. 7.

6 shows an example of a database structure in a general load balancer. According to the user's connection request, through the address search module that finds or connects to the database's allocation space from identifiers such as Internet address information, the user can save or search the allocation server information or status information for each connection. In general, each information slot is used according to a user connection request, which may require tens of millions or more of slots, which may require a very large storage space. In this case, it is very costly to construct a load balancer, and the search speed is lowered, which can be a major factor in performance degradation. Having an address by identifier for fast retrieval involves a problem that may require more storage space.

The present invention is to solve the existing problems of high cost, low performance and to realize high performance and high speed operation by configuring a database for managing connection information or session information of the load balancer in a manner suitable for the load balancer.

An object of the present invention is to efficiently use a storage space, to solve the problem of limited connection management according to a limited storage capacity, and to realize high speed operation.

The technical problem to be achieved in the present invention is summarized as follows.

-Unlimited connection management even in limited storage space.

-Significantly reduce the requirements for required storage space by realizing the effect of doubling the limited storage space

-Speed up searches in large storage spaces

-Efficient storage management that can utilize 100% of storage space.

1 is a block diagram of the present invention

Figure 2 is a database configuration using the individual military identifier slot of the present invention

3 is a block diagram of a database using a military identifier pool of the present invention

4 is a configuration diagram of a database using a separate information allocation pool of the present invention.

5 is a configuration diagram of a database using an identifier allocation method according to the server group of the present invention.

6 is an example of a conventional database method using a mass storage location.

7 is an application example of the present invention

<Explanation of Signs of Major Parts of Drawings>

11: identifier-address translation module 12: identifier information connection space

13: Information allocation space 14: Identifier information recovery block

21: identifier-address translation module 22: individual military identifier slot block

23: Individual military identifier slot

31: identifier-address translation module 32: military identifier pool block

41: identifier-address conversion module

42: military identifier management designator slot block

43: Information allocation slot pool block

51: identifier-address translation module 52: server-specific information slot block

61: address search module 62: information slot table block

63: information slot

71: load balancer central processing unit 72: server information database

73: connection information database 74: load balancing rule database

A general configuration of a database for a load balancing device proposed by the present invention is shown in FIG. 1.

The identifier-address conversion module 11 receives an identifier as input and hands over a new information allocating space, or hands over an address of an already designated storage location. Identifiers are identified by combining the Internet address, connection information, and user information of the user connection. The identifier-address translation module 11 operates in combination with a method such as a hashing algorithm to determine the identifier, the number of actual servers, the size of the storage location, the size of the military management unit, and the like. Configure to return the address value. The identifier information connection space 12 is divided according to the identifier to store the information of each connection or group of connections, or manages the connection to the information allocation space. The information allocation space 13 can be configured separately as needed, and in order to efficiently manage the storage space, the actual storage place is managed in the form of a buffer pool. The identifier information recovery block 14 is responsible for recovering the identifier information according to the identifier-to-address conversion module and the database operation mode.

The identifier information recovery block 14 includes two operation functions as follows.

First, if the user connection identifier is not stored in the database separately, the load balancer sends the data packet's sending address using the actual identifier address information when sending data to the real server, and sends the data packet as a result of the real server's processing. When the sender address is set to the real server address and the destination address is sent as the real identifier address, the load balancer receives this data packet and then converts the sender address to a virtual address assigned to the load balancer. If the real server sends this data packet directly to the real user, it can be configured not to cause this behavior.

Secondly, when the user connection identifier is stored separately in the database, when the data balancer sends the data packet to the real server, the sender sets the sender address to the appropriate address assigned to the load balancer and sends the data packet from the real server to the user. The data packet is transmitted to the original sender address and the virtual address using the stored user connection identifier. Even when storing the user connection identifier, it can be configured to operate in the first case form.

The internal structure and operation of the identifier-address translation module, the identifier information connection space, and the information allocation space may be configured by the methods shown in FIGS. 2, 3, 4, and 5, or a combination thereof. The identifier information recovery block operates according to each configuration method. The operation method is as follows.

1) Database configuration method using individual military identifier slot (Fig. 2)

2) Database configuration method using military identifier pool (Figure 3)

3) Database configuration method using separate information allocation pool (Figure 4)

4) Database configuration method using the identifier assignment method according to the server group (Fig. 5)

5) Database composition method combining 1), 2), 3), 4)

First, a method of constructing a database using individual military identifier slots is shown in FIG. The identifier-address translation module 21 passes the address of the individual group identifier slot 23 according to the identifier. The military identifier manages several identifiers in one group and manages combination of status information. The set of military identifier slots is managed by military identifier slot block 22. The data packet returned from the real server is returned by the identifier information recovery block 14 of FIG. 1 and sent to the user. Military identifiers can be controlled as appropriate depending on the size of the storage location.

Secondly, a method of constructing a database using a military identifier pool is shown in FIG. 3. The identifier-to-address conversion module 31 returns the military identifier according to the identifier of the user in a form similar to the method using the military identifier slot. The military identifier pool block 32 has several information allocation slots for each military identifier. The information allocation slots behave like individual group identifier slots in a database method using individual group identifier slots, but allow flexible allocation and use. Other operations allow the application of database methods using individual group identifier slots.

Third, a database configuration method using a separate information allocation pool is shown in FIG. This method is a method of managing flexibility by collecting information allocation slots in a separate information allocation slot pool block 43 as a method of improving flexibility in a database configuration method using a military identifier pool. The military management identifier management connector designator slot block 42 has a connection designator instead of an information allocation slot, and is connected to the military management identifier management connection designator slot.

Fourth, a method of configuring a database using an identifier allocation method according to a server group is illustrated in FIG. 5. This method is an improvement on the method of constructing a database using individual military identifier slots, which minimizes the storage location. In other words, the real server identifier is used as a military identifier. This method can be used in parallel with each method. In other words, each method can be configured to apply this method when the storage location reaches its limit.

1) Clear storage space limit: realize unlimited connection management even in limited storage space.

2) Use of storage space expansion: Significantly reduce the requirements for storage space required by realizing the effect of doubling the limited storage space

3) Speed up search: Speed up search in large storage space

4) Efficient storage management: Efficient storage management that can utilize 100% of storage space.

5) Low Cost: Reduces the cost of building a load balancer by reducing the required storage space.

6) Massive scalability: Facilitate the scalability to manage large connections.

Claims (2)

A method of configuring a load balancer database including an identifier-to-address translation module, an identifier information connection space, an information allocation space, an identifier information recovery block, a method of managing the same, and a system thereof A method and system for configuring a load balancer's database by using individual military identifier slots, using a military identifier pool, using a separate pool of information allocation, using an identifier allocation scheme by server family, or a combination thereof.