「Unbreakable」Cloud-Native Virtual Data Warehouse

TDE (Transparent Data Encryption) encryption is applied upon data storage, zero data leaks


Cloud-native serverless architectureUnafraid of downtime


TDE (Transparent Data Encryption) encryption is applied upon data storage, zero data leaks

Core Value
ReduceWarehouse Hardwareand Management Costs

Multiple data warehouses are merged into one cloud-native Data Warehouse with virtualization technology, breaking down data silos in traditional data warehouse scenarios and addressing the issue of redundant data duplication. This helps enterprises reduce the complexity of data warehouse management and achieve more flexible configuration of storage and computing resources in the cloud at lower costs.

Cost-efficientUtilization-Benefitof Data Computing Resources

By scaling data computing resources on demand, computing resources can achieve optimal allocation, enhance the agility and elasticity of the data warehouse, unlock limitless data computational capacity, better empower business growth, and move towards environmental sustainability.

「Unbreakable」| High Security, High Availability, High Reliability

PieCloudDB leverages end-to-end security technology to encrypt all data before it is written to disk, from client connection to disk encryption. It also utilizes serverless technology to take advantage of unlimited computing resources and elasticity in the cloud, ensuring the virtual data warehouse is always available. Object storage and cross-cloud disaster recovery capabilities ensure that data is never lost.

Technical Breakthrough:
Data Warehouse Virtualization
Cloud-NativeArchitecture of Separated Storage and Computing

By using a three-layer architecture of metadata-computing-data separation, independent management of storage and computing resources can be achieved in the cloud. Cloud computing resources can be elastically allocated. When there is a query or computational task, resources are launched on-demand, and costs are calculated based on usage time and resource.

New Storage Architecture "JANM"and Cache Design

At the computing layer, multiple cache layers are designed for both metadata and user data on each computing node to avoid network latency and data movement, improve computing efficiency, and ensure users' real-time requirements. PieCloudDB has also designed an efficient file format for the underlying object storage to improve computational efficiency while reducing network requests.

PieCloudDB Optimizer"Dutch": Specifically Designed for Cloud-Native and Distributed Scenarios

PieCloudDB can generate statistics more intelligently and efficiently, enabling more efficient query plans. The optimizer "Dutch" is designed based on PostgreSQL Optimizer, supports advanced features such as aggregation pushdown, pre-calculation, and block skipping, fully meeting various complex analytical query requirements.

eMPP (elastic Massive Parallel Processing)Technology

In the cloud, PieCloudDB uses the eMPP architecture to achieve concurrent execution of tasks across multiple virtual data warehouses, allowing enterprises to flexibly and rapidly scale (scale-up or scale-out) and achieve efficient scalability with changing workloads, easily handling PB-level massive data.

PieCloudDBMeta Data Management

PieCloudDB leverages the open-source database FoundationDB to implement its distributed metadata storage system called mStore. By separating the system table data and storing it in a public area, mStore can meet the needs of multiple virtual data warehouses, the on-demand creation, destruction of virtual data warehouses.

PieCloudDB Design Principle
Compliance withDB Standards
Compatible with theDW Ecosystem