EmbeddingGemma 2: an open, lightweight multimodal embedding model
deepmind.google·6d ago
AI Engineering
Hacker News·Sep 10·5 min read
Neki by PlanetScale
TL;DR
Neki, a sharded Postgres solution by PlanetScale, enables scaling beyond single-machine limits to handle hundreds of millions of queries per second and petabytes of data. It features a robust architecture that includes a router and control plane, ensuring zero downtime during scaling operations.
✦ Why It Matters
Consider adopting Neki for your Postgres workloads to achieve unprecedented scalability and reliability.
Key Takeaways
- Utilize Neki to scale PostgreSQL beyond single-machine limits.
- Implement online resharding to manage growing workloads without downtime.
- Leverage atomic transactions across shards for data consistency.
- Run schema changes seamlessly across multiple shards.
How It Works
Neki operates by creating multiple shards of a PostgreSQL database, each functioning as an independent instance. It includes a router that directs queries to the appropriate shard based on the data's location.
As workloads increase, Neki can split hot shards and create new target shards, replicating data to ensure continuity. This allows for dynamic scaling and rebalancing of data without requiring application downtime or complex maintenance windows.