Amazon Aurora
MySQL- and PostgreSQL-compatible cloud relational database service offering high performance and high availability
YDB
Scalable, fault-tolerant Distributed SQL with strict ACID guarantees
YDB delivers a distributed SQL engine with horizontal scalability, strong consistency, ACID transactions, and built-in disaster recovery, supporting both row/column tables and PostgreSQL/Kafka compatibility.
Overview
Overview
YDB is a distributed SQL database designed for interactive web services and enterprise applications that require strict consistency and full ACID guarantees. It targets developers and operators who need a single system to handle both transactional (OLTP) and analytical (OLAP) workloads while maintaining high availability.
Capabilities
The platform offers row‑oriented and column‑oriented tables, persistent queues (topics), and a rich SQL dialect (YQL). It provides PostgreSQL‑compatible and Kafka‑compatible modes, enabling seamless integration with existing tools. Fault‑tolerant architecture replicates data across nodes, racks, and datacenters, allowing reads and writes to continue during zone outages. Storage and compute layers scale independently, so you can add capacity where it is needed without disrupting service.
Deployment
YDB runs on x86_64 machines with at least 8 GB RAM, primarily on Ubuntu Linux, and can be deployed manually, with Ansible, or on Kubernetes. Multitenant and serverless configurations let multiple databases share a common pool of resources, simplifying management for SaaS platforms and variable‑load workloads.
Highlights
Horizontal scaling of storage and compute layers
Fault‑tolerant architecture surviving node, rack, or datacenter failures
ACID‑compliant transactions across multiple nodes with strict consistency
PostgreSQL‑compatible and Kafka‑compatible modes for seamless integration
Pros
- Strong consistency with full ACID guarantees
- Independent scaling of compute and storage
- Built‑in automatic disaster recovery
- Supports both OLTP and analytical workloads via row/column tables
Considerations
- Requires minimum 8 GB RAM and x86_64 hardware
- Operational complexity for multi‑zone deployments
- Primarily supported on Ubuntu Linux in production
- Learning curve for YQL and custom deployment tooling
Managed products teams compare with
When teams consider YDB, these hosted platforms usually appear on the same shortlist.
Looking for a hosted option? These are the services engineering teams benchmark against before choosing open source.
Fit guide
Great for
- Enterprises needing globally consistent, high‑throughput transactional databases
- Developers building multitenant SaaS platforms
- Applications requiring seamless scaling of compute and storage
- Workloads that benefit from PostgreSQL or Kafka compatibility
Not ideal when
- Small projects with minimal resource constraints
- Teams preferring fully managed cloud databases
- Environments limited to non‑Linux operating systems in production
- Use cases that do not require strict ACID consistency
How teams use it
Real‑time bidding platform
Handles millions of distributed transactions per second with strict consistency, ensuring accurate auction results.
Multitenant SaaS analytics
Provides isolated databases sharing storage and compute, scaling resources per tenant while maintaining ACID guarantees.
Event‑driven microservices
Uses Kafka‑compatible topics for reliable message queues, enabling fault‑tolerant data pipelines.
Hybrid OLTP/OLAP workloads
Leverages row‑oriented tables for transactions and column‑oriented tables for fast analytical queries within the same cluster.
Tech snapshot
C++48%
C29%
Python17%
Assembly3%
Go1%
Makefile1%
Frequently asked questions
What programming languages can I use with YDB?
YDB provides client libraries for C++, Go, Java, Python, and other languages; you can also interact via standard PostgreSQL drivers.
How does YDB achieve high availability?
It replicates data across multiple nodes and availability zones, allowing reads and writes to continue even if an entire zone fails.
Can YDB run in a serverless mode?
Yes, YDB supports serverless databases that share a pool of compute resources, simplifying resource management for variable workloads.
What deployment options are available?
Clusters can be deployed manually, with Ansible for bare‑metal/VMs, or via Kubernetes for containerized environments.
Is YDB compatible with existing PostgreSQL tools?
In PostgreSQL‑compatible mode, YDB accepts standard PostgreSQL wire protocol, enabling use of familiar drivers and tools.
Project at a glance
Active- Stars
- 4,661
- Watchers
- 4,661
- Forks
- 747
LicenseApache-2.0
Repo age3 years old
Last commit3 hours ago
Primary languageC++
Last synced 3 hours ago