Tongdun offers ApexDB, a Redis‑compatible distributed key‑value database designed for ultra‑low‑latency, high‑throughput decision workloads. Its Share‑Nothing architecture and built‑in operators enable sub‑millisecond compute‑and‑store queries at up to one million transactions per second per node, supporting real‑time fraud detection and risk assessment for financial and large‑scale internet services.
Funding
$100M raised to dateRaised to date based on public sources. This may differ from the amount the company actually raised and is based only on what is publicly available on the internet.
2OCMFounders
Product
Problem
Enterprises that require real-time decision-making, such as fraud detection or risk assessment, often rely on traditional key‑value stores that cannot sustain million‑level transactions per second or provide sub‑millisecond latency, especially when complex calculations must be performed alongside data retrieval. This limits the scalability and responsiveness of AI‑driven decision platforms.
Solution
Tongdun’s ApexDB is a self‑developed, distributed, multi‑threaded KV database that is compatible with the Redis protocol while employing a Share‑Nothing architecture. Each node can achieve up to a million TPS with roughly 1 ms response time. The database embeds native operators for common decision‑logic calculations, enabling compute‑and‑store requests to complete in under 1 ms without external processing. By exposing a standard Redis API, existing applications can integrate ApexDB with minimal code changes. The platform is offered as both private‑cloud and public‑cloud deployments, allowing customers to scale horizontally while maintaining low latency for high‑frequency decision workloads.
Target Audience
Primary customers are financial institutions, fintech platforms, and large‑scale internet services that run real‑time risk or fraud decision engines and need ultra‑low‑latency, high‑throughput key‑value storage.
Features
- Share‑Nothing distributed design that eliminates single points of contention and supports linear horizontal scaling
- Redis‑compatible protocol for drop‑in replacement in existing pipelines
- Built‑in high‑performance operators (e.g., aggregations, scoring functions) that execute within the storage layer in <1 ms
- Multi‑threaded engine delivering up to 1 M TPS per node with average 1 ms latency
- Flexible deployment options: on‑premises private cloud, public cloud, or hybrid configurations
- Strong consistency guarantees with automatic sharding and failover mechanisms