Key Concepts

SynxDB Cloud employs state-of-the-art cloud-native architecture. The following sections outline the key concepts in SynxDB Cloud to help you understand its components and functionalities.

Organizations

An organization is the highest-level administrative unit in SynxDB Cloud, responsible for managing resources and billing. Organization users are administrators with permission to manage the organization’s resources.

Accounts, User, and Warehouses

Organizations in SynxDB Cloud are divided into logical subdivisions called accounts. Each account serves as an isolated environment that contains its own data and compute resources, including users and warehouses.

Accounts

Each account is a distinct logical resource unit within an organization, managing its own database data and compute resources without access to those of other accounts. This structure ensures clear boundaries between accounts while enabling flexible resource management within each account. Organization users can create or delete accounts as needed. Each account is assigned a unique host address (domain name) for access, including connections via psql.

User

Users are defined at the account level, connecting to one or more databases and functioning similarly to PostgreSQL or Apache Cloudberry™ (Incubating) users. Each user represents an individual or system that interacts with databases within an account to perform operations such as querying, updating, or managing data. Users can be created, managed, and deleted via the console or the psql client, assigned roles and privileges to access data and warehouses in the account. Each user belongs exclusively to one account and is managed at the account level, ensuring clear access control and security.

Warehouses

Warehouses are compute engines within an account to process queries. Each warehouse is composed of multiple segments, which can be scaled horizontally as needed. All warehouses within a single account have access to a shared pool of data and perform computing tasks independently. Key features of warehouses include:

Resource Management: Allows pausing, resuming, and resizing on demand.
Scaling: Support changing segment counts for online horizontal scaling.

Object Storage

Object storage is a scalable and durable cloud storage solution. In SynxDB Cloud, it acts as the persistent storage layer for large datasets, enabling data loading, backups, and external table storage. Its flexible design ensures secure, cost-effective storage for both structured and semi-structured data.

Hybrid Transactional/Analytical Processing (HTAP)

SynxDB Cloud employs a hybrid transactional/analytical processing (HTAP) architecture that enables efficient handling of both transactional workloads (OLTP) and analytical workloads (OLAP) within a single database system. This architectural design allows enterprises to meet diverse business requirements on a unified platform, achieving a real-time closed loop between transactions and analytics.

Workload Isolation

To ensure optimal performance and stability for different workloads, SynxDB Cloud implements workload isolation mechanisms that physically separate transactional processing from analytical processing. This design eliminates resource contention between different types of workloads, enabling the system to deliver optimal performance for both high-concurrency transactional processing and complex analytical queries simultaneously.

Unified Storage Architecture

Despite compute resource isolation, all workloads in SynxDB Cloud share a unified storage architecture. This design ensures data consistency and accessibility, allowing transactional and analytical data to be seamlessly shared within the same data pool, supporting cross-workload data queries and analysis.

Storage Formats and Access Patterns

SynxDB Cloud supports multiple storage formats to accommodate different access patterns:

Row-oriented storage format: Suitable for frequent point queries, insertions, and updates, providing low-latency data access for transactional workloads.
Column-oriented storage format: Suitable for large-scale analytical queries, significantly improving analytical query performance and throughput through columnar compression and vectorized execution.

By flexibly selecting storage formats, users can optimize data access performance based on their specific business scenario characteristics.