Hardware and Software Resource Preparation - DingTalk Help Center

By default, YiDA Dedicated edition requires no special configuration for data persistence. If you have additional requirements, YiDA Dedicated edition supports dedicated configuration for the database and file object storage. This article describes the hardware and software resources required when configuring dedicated storage.

Database

Only the PostgreSQL database is currently supported. PostgreSQL v14 or later is recommended. The specific configuration requirements are as follows:

Item	Configuration	Description
Database location	Alibaba Cloud PostgreSQL is recommended (both RDS and ADB versions are supported).	Alibaba Cloud databases are recommended for their reliable operations, stable network, and guaranteed performance.
Database version	PostgreSQL 14 or later.	Versions earlier than 14 cannot deliver optimal performance.
Database specification	8C 16G SSD 200G or higher.	Based on a data volume of 1,000,000 records, a host specification of “8C 16G memory 200G SSD” or higher is recommended.
Database address	Provide a domain-based address rather than an IP address.	A domain name facilitates scaling and improves both configuration flexibility and service availability.
Database connection test	Use the connection test feature in YiDA dedicated configuration to verify network connectivity. Enable public network access and configure the network egress IP allowlist.	If you do not want to expose the public network, you must additionally purchase and integrate the dedicated DingTalk security gateway. Used to verify network connectivity and whether the Data Source can be initialized.
Database connection pool	Minimum of 1,000 available connections.	YiDA uses cluster deployment. To guarantee the connection count on each machine, increase the number of available server-side connections.
Connection keep-alive configuration	tcp_keepalives_idle=30 tcp_keepalives_interval=5	Increase the database connection keep-alive frequency to avoid invalid connections caused by network issues.
Database monitoring	Facilitates troubleshooting by DBAs or DevOps.	Simplifies day-to-day troubleshooting. Without it, performance issues cannot be quickly located.
Database bandwidth	A minimum of 10M dedicated to YiDA is recommended.	If business traffic is very high with frequent uploads and downloads, increase the bandwidth accordingly.

Configuration	Database capacity (single table)	Database capacity (data instance records)
4-core / 8G memory / 50G SSD	50,000	1,000,000
8-core / 16G memory / 100+ SSD	100,000	2,000,000
16-core / 32G / 100+ SSD	300,000	5,000,000
32-core / 64G / 200+ SSD	500,000	10,000,000
64-core / 128G memory / 200G+ SSD	1,000,000	20,000,000

Object Storage

Object storage supports three options: Alibaba Cloud OSS, Huawei Cloud OBS, and Inspur Cloud OSS. The object storage environment requirements are as follows:

Item	Configuration	Description
OSS type	Select Alibaba Cloud OSS, Huawei Cloud OBS, or Inspur Cloud OSS.	Three object storage types are currently supported: Alibaba Cloud OSS, Huawei Cloud OBS, and Inspur Cloud OSS.
Network configuration	Enable public network access for OSS.
Connectivity test	Use the connection test feature in YiDA dedicated configuration to verify network connectivity.	Used to verify network connectivity.
CORS settings	Configure CORS to allow the organization’s domain on YiDA. Set `Header` to `*`.	YiDA connects directly to OSS for file upload and download. CORS must be allowed for it to work.
Hotlink protection check	Check whether hotlink protection is enabled. If enabled, configure the YiDA domain allowlist. YiDA egress IPs are: • 106.11.0.0/16 • 203.119.128.0/17 • 59.82.0.0/16 • 140.205.0.0/16	Hotlink protection blocks any request not in the allowlist.
Time server check	If the organization uses a time server, verify that the time is accurate.	OSS requests have a time validity. YiDA sets it to 30s, and expired requests are invalid.

IT Resources: Storage Preparation

Local database

PostgreSQL database v14 or later.
Publicly accessible database domain, database name, account, and password.
Recommended server: 8-core 16G, 2T disk.

Local file OSS storage

Publicly accessible OSS bucket domain, access key (AK), and secret key (SK).

	Solution overview	Advantages	Applicable scenarios
Database	Solution 1: Dedicated cloud database	Delivers the high availability of a cloud database with multi-region redundancy and supports elastic auto-scaling. The YiDA platform is responsible for maintaining full cloud computing and cloud storage reliability.	Lets customers independently view and maintain the database for operational auditing. Suitable for storing business data on an independent cloud storage server, separated from YiDA’s public cloud business storage.
Database	Solution 2: On-premises IDC data center	Meets specific organization requirements by enabling local data center deployment and storage.	Suitable for organizations with specific data storage requirements whose business data fields cannot be stored in the cloud. Requires an in-house DBA team capable of maintaining the environment, with database primary-standby failover and physical disaster recovery in place.
Object storage OSS	Solution 1: Cloud OSS service	Delivers full file storage services with multi-region redundancy for high availability, and supports elastic auto-scaling.	Suitable for storing business data on an independent cloud storage server, separated from YiDA’s public cloud business storage.
Object storage OSS	Solution 2: On-premises OSS solution	Combined with private YiDA cluster deployment to meet specific organization security requirements, enabling local data center deployment and storage.	Suitable for organizations with specific data storage requirements whose business data fields cannot be stored in the cloud. Requires an in-house DBA team capable of maintaining the environment, with database primary-standby failover and physical disaster recovery in place.
Data analytics	Solution 1: Cloud-based big data analytics	Supports large-cluster data analytics. Intermediate data requires cloud storage for temporary staging.	Suitable for using cloud-based visual big data capabilities such as table association, table filtering, and data visualization.
Data analytics	Solution 2: On-premises database and data lake with local BI analytics	Provides all raw business database fields and data models across every dimension.	Suitable for organizations with a professional BI team that can perform data analytics directly against the local IDC using tools such as Tableau or Typhoon.