From Testnet to Mainnet | The Journey of zkDatabase
November 4, 2025
6 mins read
After five months—achieving a 0.0025% error rate—we hit key testnet milestones. Now, Orochi Network’s zkDatabase moves to mainnet, ushering in a high‑performance, verifiable, privacy‑preserving data layer for Web3’s Verifiable Data Economy.
zkDatabase by Orochi Network is entering a new chapter from testnet to mainnet after 5 months of testing, continuous iteration, and community-driven feedback. This milestone signals more than just product maturity—it affirms the viability of a high-performance, verifiable, privacy-preserving structured data layer designed for the demands of the Web3 and Verifiable Data Economy eras.
Why the Web3 World Needs Verifiable, Structured Data
In a decentralized future, the ability to **verify data **will become non-negotiable. Traditional databases do not natively support trustless proofs, structured verifiability, or cryptographic privacy guarantees. zkDatabase addresses this fundamental gap.
As a Zero-Knowledge-powered database layer, zkDatabase provides:
-
End-to-end data integrity
-
Verifiable execution using Merkle-based rollups
-
Scalable performance under concurrent access
-
Sub-second proof generation
-
High-level APIs for developers to interact with data securely
The Stepping Stone zkDatabase Testnet Achievements
Our testnet launched five months ago, designed to simulate real-world application use cases and to identify bottlenecks in both the database engine and the Zero-Knowledge Proof layer. The results were decisive.
Although the testnet phrase, zkDatabase has been downloaded over 9,600 times by B2B enterprises, reflecting strong developer interest and early adoption.
As zkDatabase enters mainnet, we continue to explore tailored solutions for businesses facing critical data challenges, particularly in sectors such as Real-World Asset (RWA), DeFi infrastructure, and AI/ML model integrity, where verifiable and privacy-preserving data is essential.
Performance overview:
-
Total operations: 274,757 ops/month
-
Error rate: 0.0025%
-
Throughput: WRITE 5 ops/s; READ 19 ops/s; ROLLUP 0.2 ops/s
The system demonstrated high stability, even under unpredictable data access patterns.
Benchmarking zkDatabase: Performance and Stability
Write Performance
Batch insert tests showed consistently stable throughput, regardless of scale:
The write performance remained consistent across all batch sizes, making zkDatabase well-suited for both real-time and batch-processing applications.
Query Speed
All read queries returned in under 15 milliseconds, even when executed on datasets exceeding 100,000 records. This ensures low-latency data access for real-time applications and analytics use cases.
Stress Testing
To validate concurrency and system durability, we performed tests with:
-
50 concurrent operations
-
3 separate worker threads
Results:
-
Throughput: 13 operations per second
-
Success rate: 100%
No failures or performance degradation were observed under load.
Merkle Rollup and API Layer
To support decentralized and audit-friendly use cases, zkDatabase integrates a verifiable rollup system. API performance benchmarks show:
| Component | Avg Latency |
|---|---|
| Database APIs | 318ms avg |
| Rollup Offchain (state + history APIs) | 165ms |
| Document APIs (Merkle Proof) | 78ms |
These latencies ensure predictable response times while preserving cryptographic verifiability.
Key Learnings and System Improvements
From the testnet phase, we gathered essential insights that led to key system enhancements:
-
Refined queue sequencing to prevent ID gaps
-
Improved concurrency management for parallel script execution
-
Optimization of proof aggregation for faster API responses
-
Enhanced error reporting and retry mechanisms at the API endpoint
Path to Mainnet
With the above criteria met, zkDatabase is now moving into its mainnet phase. Our phased rollout will ensure security, performance, and operational readiness.
Mainnet Launch Criteria
-
High Availability (HA) across sequencer and storage nodes
-
Stable performance with sustained throughput under load
-
Consistent API response time across diverse query types
-
Completion of third-party security reviews
-
Infrastructure provisioning for partner integrations
-
Developer tools, SDKs, and documentation finalized
What Comes Next
Designed for enterprise-grade deployment, the system has evolved beyond its initial scope on Mina Protocol and now offers multi-chain compatibility, enabling integration with a wide range of blockchain platforms.
Key advancements in this mainnet release include: enabling fine-grained data control and deterministic state transitions; multi-language SDK support, expanding beyond TypeScript to accommodate ecosystems like Rust, Python, and Go; and significant performance optimizations, including reduced proof sizes and faster proving times through improvements in the zkSNARK proving pipeline.
After 5 months of public testing, zkDatabase now meets enterprise-grade SLAs for latency and throughput. Internal benchmarks report median read latencies below 15 ms on 100k-row datasets, batch insert throughput of ~20 documents/sec (on 4 vCPU instances), and full-database state proofs generated and verified in under one second. Public mainnet access for infrastructure providers, enterprise partners, and early adopters will go live later this quarter.
Developer Resources and Community
The GitHub repository here https://github.com/orochi-network/zkDatabase —
Relevant links:
-
NPM package: https://www.npmjs.com/package/zkdb
-
Developer guide: https://docs.orochi.network/zkDatabase/introduction/
-
Testnet environment: https://testnet.zkdatabase.org
Conclusion
The transition of zkDatabase from testnet to mainnet marks a pivotal milestone, not just for Orochi Network, but for the broader evolution of verifiable data in the Web3 era.
Phase-1 of the mainnet rollout is already underway. Following **5 months **of public testnet activity, zkDatabase has achieved the performance benchmarks required for general availability. Internal testing shows a median read latency of under 15 milliseconds on datasets with 100,000 rows, while batch-insert throughput averages around 20 documents per second across various dataset sizes, tested on a single 4 vCPU instance. The proof pipeline now delivers sub-second proof generation in staging environments, with on-chain verification clocking in at approximately 500 milliseconds for full database state transitions.
These outcomes validate zkDatabase as a foundational data layer that is:
– Verifiable: Every state transition is accompanied by a Merkle root and a zero-knowledge proof, ensuring data correctness at its core.
– Consistent: The queue-sequencing gap reported during February’s testnet retrospective has been resolved and re-validated.
– Scalable: Load testing with 50 concurrent worker threads demonstrated stable throughput and zero data loss under high concurrency.
Public mainnet access for builders, infrastructure partners, and early adopters is scheduled to open later this quarter—ushering in a new era of provable infrastructure for Web3.
Explore With Us - zkDatabse
**Orochi Network, **s a proof-agnostic, Verifiable Data Infrastructure that transforms raw data into verifiable data—built for Web3, AI, DePIN, and real-world asset tokenization. With over **300K daily users, 1.5M monthly users, and 160M+ transactions, it currently powers more than 40 dApps **and blockchains. Backed by $12M in funding from the Ethereum Foundation and leading VCs, Orochi also supports a growing community of 500K+ members.
Its zkDatabase has been adopted by** 20+ blockchains,** while Orand and Orocle extend verifiability across** 49+ chains**. By combining Proof-Carrying Data with ZK systems like Halo2, zk-STARK, and Plonky3, Orochi delivers audit-grade integrity and slashes Ethereum data costs from** ~$25 to ~$0.002 per KB.**
Experience verifiable data in action - Join the zkDatabase live demo!
Book a Demo
