Real-Time Data Warehouse

Deliver the Power of Apache Doris

VeloDB is a modern real-time data warehouse for blazing-fast analytics on large-scale data, transforming batch analytics to real-time.

Start Free

Built by Apache Doris Creators

Top 1

In Doris Commits

24/7/365

Support for Doris

Top 1

Community Contribution

Apache Doris and VeloDB are Trusted By

Blog

Milestone: Apache Doris 3.0 Released

Starting from version 3.X, Apache Doris supports a compute-storage decoupled mode in addition to the compute-storage coupled mode for cluster deployment.

Read Blog

Slides

Why Choose Apache Doris over Elasticsearch

Apache Doris is increasingly replacing Elasticsearch. Here is a comparison of two and Doris use cases in observability, cybersecurity, and business data analysis.

Download Slides

Events

AWS & Apache Doris Meetup: Global Big Data & AI Ecosystem

On February 24, join keynotes from AWS, Apache Doris and talks from TrueWatch, WhaleOps, and Zilliz. Network and explore the future of data processing and AI.

EVOLUTION

Real-Time is the Future of Data Warehouse

20 years ago

10 years ago

Today

20 years ago

Traditional Data Warehouse

For small-scale analytics, on-premises with fixed architecture and limited scalability.

Key Features:

Shared Nothing Architecture
Massively Parallel Processing (MPP)
Columnar Storage

10 years ago

Cloud Data Warehouse

For big data analytics, with elastic scalability, cost efficiency, and easy operations.

Enhancements:

Cloud Native
Decoupled Compute and Storage
Elastic Scalability

Today

Real-Time Data Warehouse

Optimized for online / real-time analytics, delivering instant insights on large-scale data.

Additional Features:

Real-Time Ingestion
Blazing-Fast Analytics
Complex Data Types
Open Lakehouse SQL Engine

Data Lake 2.0

Offline / Batch Data Processing and DS / ML / AI

REAL-TIME

Industry-Leading Real-Time Performance

Real-Time Ingestion

Ensure data freshness with second-level real-time ingestion, high-performance upserts, and lightweight schema changes.

PB-scale

Data size

10 GB/s

Log write throughput

1 sec.

Data latency on 100 million logs

The Apache Doris-based logging system serves a PB-scale data size with over 99.9% availability and deliver a log write throughput of 10 GB/s, while maintaining data latency within seconds.

Blazing-Fast Analytics

Both for wide table queries and complex JOIN queries, it offers the fastest analytical performance in the world.

TOP 3

ClickBench (In Oct 2024)

3-8x

TPC-H (Doris vs. Greenplum)

TPC-DS (Doris vs. Trino)

In October 2024, Doris achieved top 3 in ClickBench. In TPC-H, Doris is 3-8x faster than Greenplum in JOIN queries. In TPC-DS, Doris lakehouse analytics outperformed Trino by 3x.

High-Concurrency Queries

Deliver sub-second query responses to thousands of users while utilizing minimal computing resources.

10 billion

Data rows per day

10,000 QPS

Query concurrency

150 ms

Query latency

...witnessed remarkable performance enhancements. Currently, it is capable of processing a daily volume of 10 billion data rows and attaining 10,000 QPS with a minimum query latency of 150ms.

MODERN

Embrace the Data Warehousing Revolution with Emerging Technologies

Cloud-Native & Multi-Cloud

Decoupled storage-compute, elastic scaling, cost efficiency and easy operations. Deploy on AWS, GCP, Azure, and more.

VeloDB, powering TrustDecision, is constructed on a cloud native architecture. With its storage and computing components decoupled, the system allows clusters to scale rapidly, which can bring a 300%~500% increase in resource efficiency.

Optimized for All Data

Native support for JSON, String and Bitmap, optimized for key use cases such as log analysis, user profiling and behavior analysis.

VeloDB's VARIANT data type for semi-structured data enables native JSON and nested data handling. It auto-analyzes JSON fields/types and uses columnar storage for frequent fields, boosting storage and analysis efficiency.

Build an Open Data Lakehouse

Serve as a Lakehouse SQL Engine. Directly run warehousing workloads on open formats like Hive and Iceberg.

Kwai replaced ClickHouse with Doris, simplifying its architecture from a separated data lake - warehouse to a unified lakehouse and achieving unified storage, high performance queries, and flexible data governance.