Create Data Box Frame

Available in VPC

Once you complete the Data Box Frame application, you can create a Data Box Frame. Data Box Frame is divided into an admin environment used by the manager who manages the Data Box Frame and an analysis environment used by analysts.

Create Data Box Frame

On the NAVER Cloud Platform console, navigate to > Services > Big Data & Analytics > Data Box Frame, and click [Create Data Box Frame] to create a new Data Box Frame by following the steps below.

1. Create admin environment

The manager of Data Box Frame can manage the data to be analyzed through the admin environment. You need to set up an admin server for admin access and a shared data store where shared data is stored. For data box frame management, you need at least 1 admin server and common data storage.

• Admin server
  • Ubuntu and CentOS can be selected as base images, and public datastores are mounted.
  • The minimum number of admin servers is 1 and you can create up to 4.
• Save shared data
  • Public data stores are provided mounted on the admin server and on Ncloud Tensorflow Server, Linux Server, and Hadoop Cluster inside each box, making it easy to share files.
  • You can use a minimum of 1 shared data storage and a maximum of 20, ranging from 500 GB to 10,000 GB. You can change capacity after creation.

• Memo
  • Write notes related to Data Box Frame.
• Infrastructure management permission
  • Through the expansion of infrastructure management permissions, you can separately manage server management permissions and network rules.

2. Create a Box environment for analysis

Analysts can analyze data provided by the Data Box Frame manager and custom data through the analysis environment. A Data Box Frame must have at least 1 data box.
In the analysis environment creation stage, you can proceed with setting the number of boxes and infrastructure. The box must contain at least 1 Connect Server, Ncloud Tensorflow Server, and NAS.

• Connect Server
  • A server with a Windows OS that allows you to connect to Cloud Hadoop and Ncloud TensorFlow Server using Putty on the Connect server. You can also use Ambari, Zeppelin, Jupyter Notebook, etc., through a web browser on a Connect server.
  • The minimum number of connect servers is 1 and you can create up to 4.
• Ncloud Tensorflow Server
  • It is a server provided with Jupyter Notebook, and you can select CPU and GPU types.
  • The minimum quantity of Ncloud Tensorflow Servers is 1, regardless of type, and you can create up to 4.
• Cloud Hadoop
  • Cloud Hadoop clusters are provided, with 1 Edge node, 2 Master nodes, and 2 to 8 Worker nodes per cluster.
  • The minimum number of Hadoop is 0 and you can create up to 2.
• Server
  • For convenient use inside the box, we provide a Linux server using CentOS or Ubuntu as an image.
  • The minimum number of servers is 0 and you can create up to 4.
• NAS
  • We use it to import files from external sources or export analysis results after external communications are blocked. The NAS comes mounted on both Cloud Hadoop, TensorFlow CPUs and TensorFlow GPUs, and Linux servers, so you can easily share files between Cloud Hadoop and Ncloud TensorFlow Server and Linux servers.
  • The minimum number of NAS is 1 and up to 4 can be used. You can assign each NAS with a capacity of 500 GB to 10,000 GB.

• SSL VPN
  • You must use SSL VPN to access the box environment.
  • You can create a minimum of 1 SSL VPN and a maximum of 3 times the number of Connect Servers.
• Memo
  • Create a memo related to the box belonging to the Data Box Frame.

3. Set box connection information

This step is to enter the name of the Data Box Frame and set the access information of the admin environment and each analysis environment (box). As account name for each server is generated automatically and you only need to set the password.

• Admin preferences
  • Data Box Frame name
  • Password of the admin server
  • Bucket name of Object Storage for export review
• Set analysis environment
  • Password of Connect Server
  Note

  When expanding the infrastructure management permissions, the ncp-admin account with administrator privileges can be granted to Connect Server. You should be cautious when using this account.

  • Password of Ncloud TensorFlow Server

4. Final confirmation

This is the step for requesting Data Box Frame creation after confirming request details in the previous step. It may take several hours or more for the Data Box Frame to be created. Once the Data Box Frame is created, you will be notified via email. To review the Data Box Frame creation details:

• Check the basic information of Data Box Frame.
• Check the admin environment of Data Box Frame.
• Check the box information of Data Box Frame.
• Review the details and click the [OK].