@msimoni: One thing I keep thinking about: with S3-like object storage as primitive, you could build a transactional database wit…
Summary
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.
Similar Articles
@jhleath: https://x.com/jhleath/status/2065408690992148698
The author explains how they built a compute platform capable of launching millions of sandboxes per second in constant time, focusing on decoupled scheduling and capacity aggregation using Cassandra and S3.
@gortron: S3 is the perfect place to store data, until you try to search it. Two months ago I launched Firn: open source vector +…
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.
@RuiTheBaker: just witnessed how mixedbread turned my bucket into a late-interaction database.
Mixedbread announces that users can now bring their own cloud bucket, enabling zero-retention indexing and search with late-interaction models.
@songhan_mit: Explore our continued efforts on KV cache compression:
A tweet from Song Han highlights continued work on KV cache compression, featuring a blog by Weian Mao that discusses system-level aspects often overlooked in papers.
@m_sirovatka: KV Cache re-use is the most important thing for agentic rollouts. We've integrated Mooncake Store into prime-rl with vL…
vLLM integrates Mooncake Store for distributed KV cache reuse, enabling cross-node prefix caching to efficiently serve agentic workloads with high token reuse.