KR20200092507A - An installation system and installation method of a distributed processing system for mass data processing applicable to various clouds and its distributed processing system - Google Patents

Abstract

The present invention relates to an installation system of a distributed processing system for processing large volume data that is applicable to various clouds, and more particularly, to an installation system of a distributed processing system for processing large volume data based on various cloud environments, an installation method thereof, and the distributed processing system. The distributed processing system that is not dependent on a specific cloud system and can be installed in various cloud environments that can be stored in various types of media by processing data at high speed at the time of processing data comprises: at least one master node installed from an installation package for the first time; at least one storage node for storing processed data; and at least one calculation node that is created through the master node when inputting data to be analyzed and processes all or a part of the data.

Description

An installation system and installation method of a distributed processing system for mass data processing applicable to various clouds and its distributed processing system}

The present invention is an installation system and installation method of a distributed processing system for processing large-capacity data applicable to various clouds, and a distributed processing system thereof. More specifically, various cloud environments are public cloud according to the location of a physical system. , Private Cloud, Hybrid Cloud, or a cloud provided by one or more commercial cloud service providers. The present invention relates to an installation system and an installation method and a distributed processing system for installing a distributed processing system for processing large-capacity data applicable to various clouds by copying an installation system to a client from an installation package that can be stored in various forms in a storage medium. .

Cloud refers to software and data connected to the Internet and stored on a central computer so that data can be used anytime, anywhere through the Internet.

Clouds come in many forms.First, depending on the type of introduction and distribution, a private cloud operated only for one group, a public cloud publicly available through an open network, and two or more clouds It can be divided into a hybrid cloud (Hybrid Cloud) with a combination that is bound together.

Currently, cloud computing technology is applied in various industries, and a cloud system is applied as a platform for processing large amounts of data. A cloud-based system can increase system availability by making necessary computing resources available at a point in time compared to existing in-house systems, and in particular, it can be more flexible in coping with rapidly changing external and internal environments because it is not dependent on in-house systems.

In addition, in a cloud computing environment, costs are incurred depending on computing resources used and time of processing. Therefore, the user pays for using the cloud computing service. However, in order to build a cloud-based distributed processing system for processing large-capacity data in a specific cloud to which the advantages of the cloud are applied, significant time and cost are incurred.

Republic of Korea Patent Publication No. 10-2014-0065545 (multi-cloud distribution management system and method, publication date: 2015.12.09) Republic of Korea Patent Publication No. 10-2018-0063240 (system and method for building, optimizing, and implementing infrastructure in a cloud-based computing environment, publication date: 2018.06.11) Republic of Korea Registered Patent No. 10-1822093 (Cloud system building device and method, registration date: 2018.01.19)

The present invention is to solve the problems of the prior art as described above, and is not dependent on a specific cloud system, and the installation system and installation method of a distributed processing system for processing large-capacity data applicable to various clouds within a short time at the time of data processing demand It is to provide the distributed processing system.

As an installation system and installation method of a distributed processing system for large-capacity data processing that is not dependent on a specific cloud system according to an embodiment of the present invention and can be applied to various clouds within a short time at the time of data processing demand, and a distributed processing system thereof,

The installation system of a distributed processing system for processing large-capacity data applicable to the various clouds includes a package including reference information files required when building and compiling the source code and binaries of the distributed processing system copied from the installation package to the client. Configuration information unit; and container creation and deletion to provide an environment for building and compiling source code and binaries for a distributed processing system for processing large amounts of data through the package configuration information unit, and building and compiling the source codes and binaries in the container. A package driver for executing; and a container having an environment for building and compiling source codes and binaries of a distributed processing system for various cloud environments; and source code and binaries of a distributed processing system are built, compiled, and generated in the container environment A control unit further comprising providing a user interface means for selecting and requesting a drive file and a cloud by a user; and storing a package that stores the drive file after the source code and binaries for the drive file are built and compiled in the container. And a external storage that can receive source code and binaries from outside.

The installation method of a distributed processing system for processing large-capacity data applicable to the various clouds is the first of the package configuration information part, the package driving part, the control part, the package storage part, and the source code and binary for the distributed processing system from the installation package. The first step, where the installation system is copied to the client,

In order to install the distributed processing system through the initial installation system, a container having an operating system environment for various cloud to be installed or an operating system environment capable of compatible operation and an environment for building and compiling is created and distributed to the package storage unit. After copying or linking the source code for the processing system and the source code and binaries of the external essential components to the corresponding container, the container creates, compiles and compresses the source code and binaries, and transmits the generated driving file to the package repository. Then delete the container. At this time, the stored drive file is a second step in which a master node drive file including a compute node drive file to be installed in various clouds and a first drive file including a storage node drive file are provided,

Step 3 to select the cloud through the user interface included in the control unit and allocate the master node and storage node in the selected cloud computing environment,

In order to provide the master system and the storage system to the master node and the storage node allocated through the selected cloud, the package driver provides network information of the master node and the storage node and other information necessary for operation in the cloud selected through the user interface of the controller. The fourth step of extracting the compressed first drive file, which is made of the configuration files of the master system and the storage system, and then copying them to a specific folder of the master drive file and the storage drive file and compressing them, respectively, to provide the second drive file. ,

The master drive file and the storage drive file of the second drive file generated in step 4 are transferred to the master node and the storage node, decompressed, and installed to provide a master system and a storage system to provide a first distributed processing system. Step 5,

When the master system and the storage system are operating normally, and when data and tasks are input from the user to the master system, a sixth step of allocating resources of the compute node and the number of compute nodes to process the data and tasks,

In step 6, the compute node driving file included in the master system is transmitted to the compute node. The transferred compute node drive file is decompressed and installation proceeds. At this time, information on whether the installation is completed can be checked through a user interface included in the control unit, and a seventh step in which a calculation system is provided,

When the calculation system is provided and the first operation, the calculation system sends a success or failure message to the master system through network information of the stored master system, and after the master system receives the calculation system success message, registers the calculation system or reinstalls the calculation system. It includes the eighth step by performing.

A distributed processing system for processing large-capacity data applicable to various clouds includes data to be analyzed from a first distributed processing system, a master system, and a user that constitutes one or more master systems and storage systems that are first installed through the installation method and installation system. And a second distributed processing system that configures one or more calculation systems generated to process all or part of the data when the task is input and one or more storage systems to store the processed data. The second distributed processing system is a distributed processing system for processing large amounts of data applicable to various clouds of the present invention.

According to the present invention, a distributed processing system for processing large-capacity data applicable to various clouds has an effect that can be operated without being limited to a specific cloud, and a separate complicated installation process is performed at a time when large data is to be processed. There is no need for customizing according to the target operating environment, and it can be installed and operated immediately.

In addition, since the distributed processing system of the present invention is provided and installed at a point in time when data analysis is required, the point and cost of eliminating the need to constantly operate the system involved in the operation of the distributed cluster system is reduced compared to the conventional techniques. There is an effect that can be done.

In addition, the above-described distributed processing system of the present invention is composed of only a master node, a storage node, and a calculation node, and does not require an additional management system, and has the effect of further reducing the operation cost of the distributed cluster.

FIG. 1 is a block diagram showing an installation system of a distributed processing system for processing large amounts of data applicable to various clouds according to an embodiment of the present invention and an overall configuration of the distributed processing system.
2 is a block diagram showing the configuration of a drive file according to an embodiment of the present invention.
3 is an overall flow chart of a method for installing a distributed processing system for processing large amounts of data applicable to various clouds according to an embodiment of the present invention.
4 is a block diagram showing an installation system configuration of a distributed processing system for processing large amounts of data applicable to various clouds according to an embodiment of the present invention.
FIG. 5 is a diagram illustrating a sequence of building and compiling packages constituting source code and binaries according to an embodiment of the present invention.
5A to 5C are views illustrating a reference information file for a package according to an embodiment of the present invention in json grammar form.
6 is a flowchart illustrating a method of generating a first drive file through an installation system according to an embodiment of the present invention.
7 is a block diagram showing a system configuration for installing the distributed processing system through a first driving file generated through an installation system according to an embodiment of the present invention.
8 to 9 are views illustrating user interface means included in a control unit according to an embodiment of the present invention.
10 is a flowchart illustrating an operation flow of a system for generating and installing a second driving file and a first distributed processing system through a first driving file generated through an installation system according to an embodiment of the present invention.
11 is a block diagram of a second distributed processing system according to an embodiment of the present invention.
12 is a flowchart illustrating an operation flow for generating a calculation system in a master system according to an embodiment of the present invention.
13 to 16 are views illustrating an embodiment in which the distributed processing system installed in various clouds is applied through an installation system according to an embodiment of the present invention.

Throughout this specification, when a part is'connected' to another part, this includes not only the case of being directly connected, but also the case of being electrically connected with another element in the middle.

In addition, when it is said that a certain part'includes' a certain component throughout the specification, this means that other components may be further included instead of excluding other components, unless otherwise stated. The term ~(below) or the level of ~ used in this specification does not mean a step for ~.

In addition, the components shown in the embodiments of the present invention are shown independently to indicate different characteristic functions, and do not mean that each component is composed of separate hardware or one software component unit. That is, for convenience of description, each component is listed as each component, and at least two components of each component may be combined to form one component, or one component may be divided into a plurality of components to perform a function. The integrated and separate embodiments of each of these components are included in the scope of the present invention without departing from the essence of the present invention.

With reference to the accompanying drawings for a preferred embodiment of the present invention will be described in detail, the already well-known technical parts will be omitted or compressed for the sake of brevity.

With reference to the accompanying drawings, a preferred embodiment of the present invention will be described in more detail.

FIG. 1 is a block diagram showing an installation system of a distributed processing system for processing large amounts of data applicable to various clouds according to an embodiment of the present invention and an overall configuration of the distributed processing system.

As illustrated in FIG. 1, the entire configuration of the present invention includes an installation package 100, a client 200, and various clouds 1000. 1 illustrates a configuration in which one installation package and one client and one or more clouds are connected, but this is only an example for description and the scope of the present invention is not limited to the number of installation packages and clients and clouds. .

First, the installation package 100 is a specific program for copying the installation system for implementing the present invention to the client 200. More specifically, it may be an application or service capable of functions such as copying, building, and compiling the source of specific software in the form of a commonly used desktop PC or server station application or web service.

The installation system is copied from the installation package 100 of the client 200, and the configuration of the installation system includes a package configuration information unit 210, a package driving unit 220, a package storage unit 230, and a control unit 240, It includes a container 250, a driving file 260, and an external storage 270.

First, the package configuration information unit 210 includes a reference information file that builds and compiles source codes and binaries for the driving file 260. At this time, the information recorded in the reference information file includes information on the package constituting the drive file 260 for installing the distributed processing system of the present invention. The above information includes location information of the package's source code and binaries, and information about essential packages that need to be built or compiled before building and compiling. At this time, the location information can include Github, ftp, http, or remote/local storage location information. In the case of a Github address, branch information can be added for installation according to the version. When is stored may additionally include additional information to confirm normal decompression.

The package driver 220 is executed from the request of the user and reads the source code and binaries for the drive file 260 to be installed in various clouds through the reference information file of the package configuration information unit 210 and stores it in the package storage unit 230 Create and delete the container 250 that provides an environment for building and compiling source code and binaries for the package, and build and compile the source code and binaries in the created container 250.

The package storage unit 230 stores source code and binaries for generating the first drive file 260, which is the drive file 260, and stores the drive file 260 in which the source code and binaries are built and compiled. do. In addition, it includes specific files such as executable files that convert network information of the master node and network information of the storage node to a specific environment configuration file allocated from the cloud selected by the user, and generates a second drive file generated through the specific file. To save.

The controller 240 makes a cloud selection and request by the user. The cloud selection and request provides a means for the user to directly select the cloud through the user interface 20 provided by the display device 10. In addition, the control unit 240 includes various cloud computing environment information.

The container 250 has an environment for building and compiling source code and binaries for the package of the driving file 260 to be created at various times and installed in various clouds. The build and compile performed in the container 250 is deleted when the build and compile are executed for the various cloud environments without environment information on the computing environment to be installed, and when the operation ends. The container 250 may be installed in the future by the package driver 220 and configured in the same way as the computing environment. At this time, the configured computing environment physically occupies a specific space in the client 200, but is logical, system The enemy is an isolated environment.

The driving file 260 is for installing the distributed processing system of the present invention, and the source code and binaries of the package for the driving file are the results of the build and compilation, and are stored in the package storage. The drive file 260 includes a master node drive file 3000, a storage node drive file 4000, and a compute node drive file 5000. In addition, the driving file 260 may exist as a compressed file and may be stored in various storage media. The corresponding drive file 260, the master node drive file 4000, the storage node drive file 5000, and the compute node drive file 6000 are each installed in the assigned node and contain only necessary configuration information when interconnected. not.

The external storage 270 is connected and used outside the client 200. The external storage 270 builds by storing the source code and binaries in the external storage 270 in the package storage 230 when the source code and binaries for the package requiring compilation and building are not in the package storage 230 And help compile. External storage can be in the form of websites, programs, etc.

The first drive file 260a including the master node drive file 3000, the storage node drive file 4000, and the compute node drive file 5000 is generated through the installation system including the above configuration, and the result is the The first driving file 260a is generated and stored in the package storage unit 230 which is an installation system of the client 200.

Thereafter, the installation package 100 for the implementation of the present invention at a user's need is not a single operation program or system, but a distributed processing system capable of performing a specific function after installing one or more programs or systems and connecting them to each other. It has additional features to install.

In the distributed processing system to be installed in various clouds through the installation system, the cloud environment is a public cloud, a private cloud, a hybrid cloud, or one or more types depending on the physical system location. It may include a cloud provided by a commercial cloud service provider. The distributed processing system includes one or more master systems 300 to be initially installed in the cloud through an installation system, and one or more calculation systems that process all or part of data from the master system 300 when data to be analyzed is input ( 500) and one or more storage systems 400 for storing processed data.

The master system 300 is a part of the distributed processing system of the present invention that is first installed in the cloud through the first drive file 260a generated from the installation system, and the master node drive file 3000 which is the first drive file 260a ) Includes the compute node drive file 5000 and allocates the compute node 500 by calculating the number of compute nodes and the resources of the compute node when data and tasks are inputted by user requests after being installed in the node allocated to the cloud. do.

The storage system 400 is transmitted and installed when the master node driving file 3000 is transmitted to the cloud. The installed storage system 4000 includes data processed by the calculation system 500 and information necessary for data analysis.

The calculation system 500 is generated from the master system 300 when the master system 300 and the storage system 400 operate normally and data and tasks are input from a user. At the time when the calculation system 500 is installed, it includes network information of the master system 300. In addition, the network information of the own calculation system 300 installed in the master system 300 is registered and connected through the network information of the corresponding master system 300.

2 is a block diagram showing the configuration of the drive file 260 of the present invention.

The drive system 260 for installing a large-capacity data distribution processing system applicable to various clouds of the present invention includes a master node drive file 3000, a compute node drive file 4000, and a storage node drive file 5000, The first driving file 260 is generated by building and compiling the source code and binary of the package for the driving file 260 in the installation system container 250. Also, the first generated driving file 260 may be defined as a first driving file 260a. The master node drive file 3000 of the drive file 260 includes a calculation node drive file 5000 and is composed of systems and information for installing other calculation systems 500. The compute node driving file 5000 is included as the master system 300 and is installed as a system for processing data and tasks on nodes allocated to the selected cloud. The storage node driving file 4000 is installed as a system capable of storing data to be analyzed and data required for analysis in the calculation system 600.

The driving file 260 may be defined as a first driving file 260a and a second driving file 260b according to information and forms included as the distributed processing system is installed according to an embodiment.

3 is an overall flow diagram of a method of installing a distributed processing system for processing large amounts of data applicable to various clouds in an embodiment of the present invention.

In a method of installing a distributed processing system for processing large-capacity data applicable to various clouds, the installation system is first copied from the installation package 100 to the client 200. (S100)

The package storage unit 230 of the installation system copied from the installation package 100 is source code for a master system 300, a storage system 400, a calculation system 500, and other components of the distributed processing system. And source and binary of external essential components. Create a container 250 suitable for an operating system environment or compatible operating system environment of various clouds in which the distributed processing system is to be installed, and copy the source code of the client 200 and sources and binaries of external essential components to the corresponding container ( After copying or connecting to 250), build, compile, and compress it, store it in the package storage unit 230, and then delete the corresponding container 250. The master node 301 allocated to the cloud and the master node driving file 3000 for installation on the storage node 401 and the first driving file 260a which is the storage node driving file 4000 are generated to generate a package storage unit 230 ).(S200)

It provides a means for selecting a cloud through a user interface 20 provided to a user and allocating nodes to serve as the master node 300 and the storage node 400 in the selected cloud computing environment. (S300)

In order to install a first drive file 260a including a master node drive file 3000 to a node allocated to each role to be installed in the cloud, the first drive file 260a is a master node assigned to the cloud. The network information and the network information of the storage node are made into a configuration file to be used, extracted, and then the corresponding configuration file is injected to generate a second drive file. At this time, the network information injected into the master node drive file 3000, the storage node drive file 4000, and the compute node drive file 5000 may include a master node IP, a master node network interface, and a master node SSH port information. Yes (S400)

The master node 301 and the storage node 401 allocated in the above method are the master node drive file 3000 which is the second drive file 260b compressed by injecting executable files and environment setting files to install in the selected cloud. ) And the storage node driving file 4000 respectively. The first distributed processing system is provided by checking the service registration and normal operation of the master node driving file 3000 and the storage node driving file 4000, which are the second driving files extracted from the assigned node, from each node.( S500)

When the master system 300 and the storage system 400 are operating normally, and data and tasks are input from a user, the calculation node 501 is configured so that the master system 400 can properly process the input data and tasks. The number of the calculation node 501 to serve as the resource and the calculation system 500 is allocated. (S600)

After allocating the calculation node 501n through the above method, the calculation system 500 is determined by determining the calculation node driving file 5000 from the master system 300 and transmitting the calculation node driving file 5000 to the calculation node 501n. Register and check normal operation. At this time, the calculated node driving file 5000 includes network information of the master system 300. (S700)

When the calculation system 500 is installed and in operation, registering the calculation system 600 information to the master system 300 through the network information of the master system 300 in which the calculation system 500 information is stored for the first time. I can do it.(S800)

4 is a diagram illustrating an installation system of a distributed processing system applicable to various clouds in an embodiment of the present invention.

5 to 5B illustrate the order in which packages for the master node drive file 3000 of the drive file 260 according to an embodiment of the present invention are built and compiled in the container 250, and reference information files for the packages. It is a diagram shown in the form of a script file.

Referring to FIG. 5, A, B, and C marked as objects refer to Package A, Package B, and Package C constituting the master node driving file 3000. The packages constituting the master node driving file 3000 read a reference information file for the package through programs and commands. After reading, the source code and binary for the package can be built and compiled in the order of Package C, Package B, and Package A through the reference information file.

5A to 5C are diagrams illustrating reference information file information for a package according to an example.

Referring to FIG. 5A, the information configuration of the reference information file includes package name, required package name, location information, version information, executable file location information, and container file location information.

First, the package name indicates the package name for the reference information file.

The required package name indicates the package that must be preceded before building and compiling the package.

The location information displays location information about the source code and binaries for the package.

The location information for the package may be displayed as being divided into inside and outside.

The version information indicates the version of the package.

The executable file location information displays information about the location in the file for executing the container 250.

The container file location information indicates the location of a file for configuring the container 250 in a certain environment with source codes and binaries for the package.

Package B and Package C may also be configured as follows through the reference information file for Package A.

6 illustrates the overall operation flow for generating the first driving file 260 in FIG. 6 with reference to the drawings illustrated in FIGS. 4 to 5B. Also, the overall operation flow diagram shown in FIG. 6 is included in steps S100 to S200 of FIG. 3.

First, when the package driver 220, which is the installation system S100 copied from the installation package 100 to the client 200, is executed by a user request (S10), the package driver 220 receives the package configuration information 210. Read the reference information files for Package A, Package B, and Package C constituting the master node driving file 3000 (S11).

When the required package B exists in the reference information file of the package A for generating the master node drive file 3000, which is the first drive file 260a, prior build and compilation are first required through the package driver 220 again. Read the reference information file including the source code and binary location information for the required package B. At this time, if the required package continues to exist in the reference information file of package B, it proceeds in the form of recursive or loop loop through the above method to read package C. After reading the reference information file of package C, if the required package does not exist in the reference information file of package C, the recursive syntax or loop syntax ends and the package storage unit 230 uses the source code and binary for the required package. Prepare to build and compile. At this time, if the source code and the binary do not exist in the package storage 230, it is downloaded from the external storage 300 and stored in the package storage 230. (S12)

If there is no preceding required package in the reference information file of the package C, the container file for building and compiling the source code and binary is generated by reading the container file location information in the reference information file. Thereafter, in order to build and compile the container 250, the location information of the executable file is read in the reference information file, and the executable file stored in the package storage unit 230 is connected to the container 250. At this time, the package storage unit 230 and the container 250 are connected. (S13)

When the container 250 is ready, the reference information file for the package C is read to read the location information for the source code and binary. Location information is divided into internal and external location information. The internal location information includes information that can be stored in the client 200 including the package storage unit 230. The external location information is downloaded from the external storage 270 and an external storage medium, and the package storage unit 230 ). Subsequently, when the package storage unit 230 is ready, the source code and the binary for the package C are transferred to the container 250 (S14), and compilation and build are performed through the executable file stored in the package storage unit 230. .(S15)

Thereafter, the compiled and built package C is stored in the package storage 230. Thereafter, the container 250, which has been built and compiled, is deleted. (S16)

The package C has been built and compiled, stored in the package storage unit 230, and the container 250 is deleted, and then the source code and the binary for the package B constituting the master node driving file 3000 are recursive syntax or loop syntax. Build and compile through. Package A also proceeds through the recursive or loop syntax after the package B is compiled and built. When the package A, the package B, and the package C have been built and compiled, they are stored in the package storage unit 230 as a master node driving file 3000. The storage node drive file 4000 or the compute node drive file 5000 is generated through the same steps as the master node drive file 3000. When the master node drive file 3000, the storage node drive file 4000, and the compute node drive file 5000 are completed and there are no more packages to build and compile (S17), constituting the distributed processing system of the present invention The first driving file 260a is generated (S18). (S200)

7 is a diagram illustrating a system configuration for installing the first distributed processing system 260c through the first drive file 260a generated by the installation system in one embodiment of the present invention.

8 to 9 are views showing examples of the user interface 20 of the present invention.

8 is a means for a first user interface 20a for selecting a cloud.

The method for selecting the cloud 1000 through the first user interface 20a includes: a first user interface that is a selection list 30 in which a plurality of clouds 1000 are displayed as selection objects through the display device 10 ( 20a). In addition, the selection list 30 may have a form of a first user interface 20a in which at least one selection means among radio buttons, file selection formats, texts, and the like is provided to the user. The means for selecting the cloud shown in FIG. 8 is one of the selection means. Thereafter, when the selection of the cloud 1000 is completed, a second user interface 20b for checking and correcting information about a node to be used in the selected cloud is provided.

9 is a second user interface 20b means for checking and correcting information about the assigned node.

When the step of selecting the cloud 1000 is completed, information on a node allocated from the selected cloud is provided through the second user interface 20b. Information about the assigned node may be displayed in a plurality through the second user interface 20b, node setting 40 displaying information about the node assigned to the selected cloud 1000 and the environment for the assigned node. And an installation button 50 for checking information and installing the second driving file 260b generated through the first user interface 20a to the assigned node. The installation button 50 may be configured in the second user interface 20b as in the node setting 40.

Also, the second user interface 20b may be displayed and configured in an area of the first user interface 20a for selecting a cloud according to an embodiment.

10 is a flowchart illustrating an operation flow of a system for installing a first distributed processing system 260c in a cloud selected through a first drive file 260a generated by an installation system in an embodiment of the present invention. to be.

FIG. 10 describes the overall operation flow of FIG. 10 with reference to FIGS. 7 to 9, and is included as steps S300 to S500 of FIG. 3.

In order to install the first drive file 260a including the master node drive file 3000 generated through the installation system in various clouds in step S200 in various clouds, the installation system control unit 240 is a first user provided in the display device A node to serve as a master node 300 and a storage node 400 in the selected cloud computing environment by selecting a cloud by request to the user through a selection list 30 displaying a plurality of clouds in the interface 20a (S20). Is allocated through the API provided by the selected cloud. At this time, the allocated node is the first drive file 260a including the master node drive file 3000 and is allocated to the cloud selected according to the optimized resource specification. Thereafter, the control unit 240 receives the computing environment and network information for the selected cloud. (S21) (S300)

The transmitted information is created as an environment setting file according to each format in the first drive file 260a stored in the package storage unit 230 through the package driver 220, and the created environment setting file is the first drive file When the 260a is decompressed and made into each folder, it is moved to a specific folder (S22) and compressed again to generate the second drive file 260b. (S23) (S400) At this time, the first drive file 260a is The configuration file injected into the master node drive file 3000, the storage node drive file 4000, and the compute node drive file 5000 may include a master node IP, a master node network interface, and a master node SSH port information. have.

In step S23, the second drive file 260b compressed by injecting the environment setting files is stored in the package storage 230, and the master node drive file 3000 and the storage node drive file as the second drive file 260b The 4000 is a master node drive file 3000 and a storage node drive file 4000 which are the second drive file 20b to the node allocated to the selected cloud through the means of the installation button 50 of the second user interface 20b. Is transmitted. (S24)

The master node driving file 3000 and the storage node driving file 4000, which are the second driving file 260b, are decompressed through the execution file for the second driving file 260b stored in the package storage 230 (S25) ), and then the first distributed processing system 260c is generated. (S26)

The first distributed processing system, the master system 300 and the storage system 4000, provide a message through a third user interface to allow the user to confirm service registration and normal operation. At this time, when the normal operation and verification are completed, the master node 300 and the storage node 400 are connected through network information on the master node 300 included in the storage node 400. (S27) (S500)

11 to 12 are views illustrating the distributed processing system and the operation flow of the distributed processing system applicable to various clouds in one embodiment of the present invention.

As shown in FIG. 11, the distributed processing system includes a master node 300, a storage node 400, and a compute node 500 in a cloud selected by a user.

FIG. 12 shows the overall operation flow of the distributed processing system applicable to various clouds, and is included in steps S600 to S800 of FIG. 3.

In step S500, the master system 300 and the storage system 400 are operating normally, and when data and tasks are input from the user to the master system 300 (S30), the master system 300 inputs data and tasks. Calculate the number of nodes and resources of the calculation node 501 so that it can process properly. (S31) (S600)

After allocating the calculation node 501 through the calculated number of calculation nodes 501 and resources (S32), the calculation node driving file 5000 is transmitted from the master system 300. (S33) At this time, the calculation node driving file When an error occurs in the compute node 501 to which 5000 is to be allocated, a continuous request is made from the master system 300. If an error occurs even in the continuous request, another compute node 501n is allocated from the master system 300 to transmit the compute node drive file 5000. If the transfer is successful, the compute node drive file 5000 is transmitted to the compute node 501, and registration and normal operation are confirmed. (S34) (S700) The master system 300 is installed in the compute node drive file 5000 installed at this time. ) Network information.

Through the above steps, the calculation system 500 is generated (S35), and the first calculation system 500 information stored in the calculation system 500 is stored through the network information of the master system 300 to the master system 300. It may include the step of registering (S36). (S800)

13 to 16 are views illustrating an embodiment in which a distributed processing system installed in a cloud through an installation system is applied to various clouds 1000 according to an embodiment of the present invention. 13 to 16, M is a master system 300, S is a storage system 400, and W is defined as a calculation system 500.

13 is an embodiment in which the public cloud 1000a of a distributed processing system applicable to various clouds according to the present invention is applied, and accordingly, when the master system 300 installed from the installation system of the present invention inputs data and tasks In the master system 300 installs a calculation system 500 for analyzing data and tasks. In addition, resources can be scaled in and out (SCLALE IN/OUT) as many nodes as necessary.

14 is an embodiment in which a private cloud 1000b of a distributed processing system applicable to various clouds according to the present invention is applied, and typically, the private cloud 1000b has limited computing resources. When installed in the private cloud 1000b, the distributed processing system including the master system 300, the storage system 400, and the computing system 500 may be installed in the private cloud 1000b in accordance with the computing resources initially provided.

15 to 16 are views illustrating an embodiment of a multi-cloud and a hybrid cloud of a distributed processing system applicable to various clouds according to the present invention.

Hybrid cloud typically refers to a cloud having a combination of two or more clouds.

15 is a calculation system when a distributed processing system installed through the installation system of the present invention is installed in the two or more various clouds (1000a, 1000b), respectively, the master system 300, the storage system 400 is installed and data and tasks are input. An example of using a hybrid cloud environment in which 500 is installed is shown.

16 shows that the distributed processing system installed through the installation system of the present invention is the first master system 300 and the storage system 400 installed in at least one cloud 1000a through the user interface of the control unit, and the user interface of the control unit is installed. Through another cloud 1000b, the number of nodes and computing resources are extended from the master system 300, and the computing system 500 is installed in another cloud 1000b. The installed calculation system 500 is connected to the master system 300 through network information for the master system 300, and a distributed processing system is provided. The cloud 1000a in which the master system 300 and the storage system 400 are installed may be at least one of a public cloud environment and a private cloud environment.

10: display device 260n: drive file
20: user interface 270: external storage
30: selection list 300: master system
40: node setting 301: master node
50: installation button 400: storage system
100: installation package 401: storage node
200: client 500: calculation system
210: package configuration information unit 501: compute node
220: package driver 3000: master node drive file
230: package storage unit 4000: storage node drive file
240: control unit 5000: calculation node driving file
250: container

Claims (29)

A package configuration information unit including a reference information file for building and compiling the source code and binaries of the package copied from the installation package to the client;
A package driver for creating and deleting containers to build and compile through the package configuration information unit; and
Containers that build and compile for various clouds and are created and deleted by the package driver; and
A driving file generated by building and compiling the source code and binary of the package in the container; and
A control unit for selecting and requesting a cloud for installing the drive file; and
A package storage unit that stores source code and binaries for the package and a driver file that has been built and compiled; and
An installation system of a distributed processing system for processing large-capacity data applicable to various clouds, including; an external storage that can receive source code and binaries from outside, in addition to the package storage unit. According to claim 1,

The installation package, a large data processing applicable to various clouds characterized by a specific program or service capable of copying, building, compiling, etc. of a specific software source in the form of an application or web service of a desktop PC or server station Installation system of distributed processing system.

According to claim 2,

The installation package is for a large amount of data processing applicable to various clouds, which is characterized by installing a single operation program or system, installing one or more programs or systems, connecting them to each other, and installing a distributed processing system capable of performing a specific function. Installation system of distributed processing system.

According to claim 1,

The information in the reference information file includes a prerequisite package name, location information for source codes and binaries for the package, location information for executable files, version information, and container information file location information. Installation system of distributed processing system for data processing.

According to claim 1,

The package storage unit is a system for installing a distributed processing system for processing large-capacity data applicable to various clouds having a characteristic of receiving and storing source codes and binaries from a plurality of executable files and external storage.

According to claim 1,

The control unit provides a user interface for selecting various clouds and an installation system of a distributed processing system for processing large amounts of data applicable to various clouds having characteristics that include environment information for the various clouds.

According to claim 1,

The drive file is a result of an installation system for installing a distributed processing system in various clouds, and includes a master node drive file, a storage node drive file, and a compute node drive file for processing large amounts of data applicable to various clouds. Installation system of distributed processing system.

According to claim 1,

The external storage is an installation system of a distributed processing system for processing large amounts of data applicable to various clouds, characterized in that it exists in various media such as websites and programs.

According to claim 1,

The drive file is stored in various storage media and can exist as a compressed file. It is transmitted to a distributed node to be installed in various clouds and processes large data applicable to various clouds that do not contain only environment setting information required when connected after installation. Installation system of distributed processing system.

A method for installing a distributed processing system for processing large amounts of data applicable to various clouds, the method comprising:

Copying the installation system from the installation package to the client;

Generating a first drive file including a master node drive file, a storage node drive file, and a compute node drive file through the installation system;

Selecting a cloud by a user and assigning a master node and a storage node in the selected cloud computing environment;

Generating a second drive file by injecting an environment setting file into the first drive file to install a first drive file in the selected cloud;

Transmitting and decompressing the second drive file, the master node drive file and the storage node drive file, at each node to confirm a first distributed processing system registration and normal operation;

Through the step of transmitting and registering the second drive file, the master system and the storage system are operating normally, and when data and tasks are input from the user to the master system, the resources of the compute node and the node to act as the compute node Assigning a number;

Determining the compute node drive file from the master system after allocating the compute node, registering the compute node drive file to the compute node, verifying normal operation, and including the network information of the master system in the compute node drive file installed;

A method of installing a distributed processing system for processing large-capacity data applicable to various clouds, including the step of determining and registering the calculation node driving file and configuring a second distributed processing system.

The method of claim 10,

The generating of the first driving file may include reading a reference information file for a package constituting the driving file through the package driver through a user request;

Checking whether there is an essential package that must be built and compiled first in order to have the driving file;

Creating a container through a package driver to build and compile a required package in advance, and then reading a reference information file for the package to connect an executable file and a container in the package storage unit to connect the container and the package storage unit;

Reading the location information of the source code and the binary for the package from the reference information file and transmitting the source code and the binary to a container,

Building and compiling the source code and the binary for the package in a container through the step of transmitting the source code and the binary to a container,

When the compile and build is completed, the completed package is stored in the package storage unit, and the container is deleted.

If there are source codes and packages that have not been compiled and built, creating a container again to execute the build and compile as recursive or loop statements,

Generating a first driving file through the step of building and compiling the source code and the binaries; A method of installing a distributed processing system for processing large amounts of data applicable to various clouds, including.

The method of claim 11,

The step of checking the presence or absence of the required package is performed by checking necessary essential packages in advance to generate the drive file through a reference information file for a package constituting a drive file included in the package configuration information part through the package driver, and A method of installing a distributed processing system for processing large amounts of data applicable to various clouds having a characteristic of compiling.

The method of claim 12,

In the process of reading the reference information file for the package constituting the driving file, if there is a required package for the package, it is applicable for processing large amounts of data applicable to various clouds having characteristics that are dependent on loop or recursive syntax. How to install a distributed processing system.

The method of claim 11,

In the generating of the container, a distributed processing system for processing large-capacity data applicable to various clouds having characteristics that are appropriately generated for various clouds by creating containers from the package driving unit through environmental information on various clouds included in the control unit How to install.

The method of claim 11,

In the step of transmitting the source code and binary to the container, when there is no source code and binary in the package storage inside when referring to the internal location information of the source code and binary through the reference information file, the source through an external storage A method of installing a distributed processing system for processing large amounts of data applicable to various clouds having a feature of storing codes and binaries in the package storage unit.

The method of claim 10,

The cloud selection and request step is distributed processing for processing of large-capacity data applicable to various clouds including a selection list means for selecting a cloud to be used among a plurality of clouds by receiving a first user interface from a control unit How to install the system.

The method of claim 16,

The selection list is a radio button, a file object selection, a method for installing a distributed processing system for processing large amounts of data applicable to various clouds, characterized in that it has a first user interface type having at least one or more selection means among text.

The method of claim 10,

In the step of allocating the master node and the storage node to the selected cloud, the second user interface for allocating the node after the cloud selection and request step through the control unit may be displayed as a plurality of user interfaces, and allocating the node to the selected cloud A method of installing a distributed processing system for processing large amounts of data applicable to various clouds, including a node setting and completion button for entering and proceeding with the environment for the node to be allocated.

The method of claim 10,

The step of generating a second drive file by injecting an environment setting file to the first drive file is generated as a plurality of first drive files suitable for various cloud environments, and then environmental information for the cloud selected through a user interface to the control unit To create a configuration file in a specific folder and store it in the package storage section through the package driver, decompressing the first drive file and injecting the generated configuration file to apply to various clouds generating a second drive file Installation method of distributed processing system for processing large data as much as possible.

The method of claim 10,

In the step of registering the first distributed processing system as a service, the generated second driving file is transmitted to a node assigned to a selected cloud through a specific file including an execution command stored in a package storage unit, and decompressed to register a service. And a method of installing a distributed processing system for processing large amounts of data applicable to various clouds to check normal operation.

The method of claim 20,

The step of confirming the normal operation of the first distributed processing system service is for processing a large amount of data applicable to various clouds that provide a third user interface through a control unit so that a user directly checks a message to confirm normal operation and service registration. How to install a distributed processing system.

The method of claim 10,

In the step of registering the second distributed processing system, a large-capacity data applicable to various clouds that allocates a compute node from a master system and decompresses the compute node-driven file by transmitting it to the assigned compute node to check service registration and normal operation How to install a distributed processing system for processing.

The method of claim 22,

When transmitting the compute node driving file to the compute node allocated from the master system, if a transmission error occurs, a continuous request is made from the master system or a new compute node is created and transmitted from the master system, which is applicable to various clouds. A method of installing a distributed processing system for processing large amounts of data.

The method of claim 22,

The transmitted calculation node driving file includes network information of the master system, and when installed in the calculation node and provided with the calculation system, has a feature of transmitting its calculation system information to the master system through the network information of the master system. A method of installing a distributed processing system for processing large amounts of data applicable to various clouds.

One or more master systems to be initially installed in the cloud through the installation system; and one or more calculation systems that process all or part of the data through the master system when data to be analyzed is input; and one or more storage to store processed data Large data distribution processing system applicable to various clouds, characterized by including;

The method of claim 20,

The master system includes a calculation node driving file, and when data and tasks are input, a massive data distribution processing system applicable to various clouds having a feature of calculating and allocating the number of calculation nodes and resources of the calculation node from the master node .

The method of claim 20,

The storage system is a large-scale data distribution processing system applicable to various clouds, characterized in that it includes information required for data analysis and data processed by the compute node.

The method of claim 20,

The calculation system includes network information on the master node at the time of installation, and is applied to various clouds characterized by registering information on the calculation node on the master node through network information on the master node. Large data distribution processing system possible.

The method of claim 25

The distributed processing system, which is installed in various clouds through an installation system, has a characteristic applied to various clouds provided by public cloud, private cloud, hybrid cloud, or one or more commercial cloud service providers depending on the location of the physical system. Large data distribution processing system applicable to various clouds.

Images