Tag
A technical comparison between ZeroFS and Amazon S3 Files, two systems that expose POSIX filesystems backed by object storage. The article highlights differences in storage layout, object interoperability, and trade-offs between direct S3 access and ZeroFS's packing, compression, and encryption.
Firn is an open-source, multi-tenant vector and full-text search engine backed by object storage like AWS S3, providing a tiered storage architecture with RAM and NVMe caching for performance. It achieves sub-second cold queries with IVF_PQ indexes and microsecond warm hits via result caching.
The TiDB Cloud team built an AI Agent cloud disk called drive9.ai based on TiDB, aiming to provide a queryable, programmable, and governable file system layer for Agents, solving engineering pain points such as file and metadata consistency, object storage latency, and file understanding in AI scenarios.
The author demonstrates how to use the billy filesystem abstraction to make a Tigris object storage bucket behave like a filesystem, enabling go-git to treat it as a git repository server directly.
Mixedbread announces that users can now bring their own cloud bucket, enabling zero-retention indexing and search with late-interaction models.
LangChain engineers detail how they built a custom inverted index for SmithDB to support full-text search and JSON filtering over large agent traces stored in object storage, achieving median latency of 400ms despite huge payload sizes.
A tweet discusses the idea of building a transactional database with infinite throughput using S3-like object storage and content-addressing, where blocks are written in parallel and the root hash is updated periodically.
ObjectCache proposes using S3-compatible object storage for LLM KV cache reuse to reduce cost and increase capacity, with a co-designed storage protocol and transfer schedule that minimizes latency overhead. Experiments show it adds only 5.6% latency over local DRAM for 64K contexts.
Gergely Orosz announces turbopuffer as a season sponsor for The Pragmatic Engineer Podcast, highlighting the database's innovative use of object storage and smart caching that helped companies like Cursor scale while cutting costs by 95%.
Hugging Face launches Storage Buckets, a purpose-built storage solution for AI teams offering per-TB pricing, built-in CDN, and Xet deduplication for model weights, datasets, and checkpoints.
Files SDK is a unified storage SDK for object and blob backends, simplifying integration across different storage providers.