October 2025 Summaries

PlanetScale has introduced a new pricing tier for its Postgres databases, offering a single-node, non-high-availability (HA) mode at $5 per month, aimed at developers and startups who require quality and features like Insights without the cost of a multi-node configuration. This new option, perfect for development, testing, and non-critical workloads, allows customers to vertically scale their single node as needed, maintaining durability without additional replicas. The new PS-5 node type provides an affordable entry point, offering flexibility for businesses to start small and handle rapid growth without complex migrations, ensuring they can eventually scale to larger, more robust configurations without changing their database platform.

The benchmarking exercise conducted by Ben Dicken evaluates the performance improvements in Postgres 18 compared to Postgres 17, focusing on the new io_method configuration option, which introduces io_uring for asynchronous disk reads alongside the traditional sync and new worker methods. The benchmarks, run on various AWS EC2 instance configurations with differing IOPS and storage types, reveal that while Postgres 18 shows notable improvements in I/O handling, especially with the worker method as the new default, the anticipated performance boost from io_uring is not as significant across all scenarios. The tests demonstrate that local NVMe storage consistently outperforms network-attached options, highlighting the importance of low-latency, high-IOPS environments for optimal database performance. Despite io_uring's potential, it is shown to excel mainly under high I/O concurrency, and does not universally outperform other methods, emphasizing that no single I/O configuration is best across all workloads.

With the rise of modern embedding models transforming data into multi-dimensional vectors, the integration of vector indexes into relational databases has become essential, yet challenging due to the absence of existing research on implementing vector indexes within these databases. The article discusses the integration of Hierarchical Navigable Small Worlds (HNSW) into MySQL for PlanetScale, emphasizing the necessity of supporting larger-than-RAM indexes for real-world applications, as relational databases typically manage data that exceeds available memory. The design addresses limitations of HNSW, such as its static nature and lack of transactional support, by developing a hybrid vector index that combines in-memory HNSW for performance and on-disk storage for scalability. This hybrid approach maintains transactional integrity and crash resilience through a Write Ahead Log (WAL) while optimizing for inserts, updates, and deletes. The article highlights the implementation of background maintenance operations, such as splits, reassignments, and merges, to ensure the index remains efficient and accurate, offering a promising solution for vector indexing in relational databases.