EloqDoc is flipping the script on document databases. EloqDoc couples MongoDB-style document models with a bold durability bet: object storage like S3 is the first-class durability layer, while fast local caching is just the acceleration layer. That combo aims to cut cost and boost resilience without sacrificing compatibility [1].
Durability as First Citizen — In practice, object storage replaces traditional block storage as the primary durability surface. A local cache of NVMe drives keeps hot data fast, so you get quick reads without paying for always-on hot block storage [1].
Scale, Distribution & Multi-Writer — EloqDoc is described as a native distributed system with true multi-writer support, removing the need for manual sharding routers like mongos while keeping full MongoDB driver compatibility [1].
Performance vs. Cost — The project highlights auto-tiering: a 5TB+ database can live mostly on cheap, durable object storage (with cross-AZ replication) and stay expensive-free on spinning down data movement. The setup touts S3-level durability at lower cost, plus NVMe-backed hot paths delivering fast IOPS, thanks to the local cache layer [1]. It also notes that object storage can be up to 3x cheaper than traditional cloud block storage like EBS [1].
Compatibility & Deployment — EloqDoc remains fully compatible with existing MongoDB clients and drivers, aiming to merge familiar workflows with a durability-first architecture [1].
Closing thought: EloqDoc’s approach invites a rethink of reliability and cost in MongoDB-like workloads—watch how the decoupled compute/storage and multi-writer design lands in real-world apps [1].
References
Show HN: EloqDoc: MongoDB-compatible doc DB with object storage as first citizen
Open-source document DB on Data Substrate uses object storage as primary durability; MongoDB-compatible with decoupled compute and multi-writer capabilities.
View source