SShortSingh.
Back to feed

Usai Runtime Isolates Request State to Prevent Cross-Request Data Leaks

0
·4 views

Most backend servers run inside a single long-lived process, meaning mutable state created during one request persists and can bleed into subsequent ones. This behaviour, common across Node.js, AWS Lambda, and Cloudflare Workers, forces developers to manually manage state isolation — a discipline that often fails silently in production. Usai, an alpha-stage runtime currently at version 0.0.10, takes a different approach by creating a fresh execution environment for each unit of work, so module-level variables reset on every request. Benchmarks show that while this isolation carries a CPU overhead of 6–14 times that of Node on trivial routes, Usai matches Node throughput on PostgreSQL-touching routes. The project's core argument is that persistent infrastructure — connections, listeners, compiled code — need not imply persistent application state, and that the boundary between the two should be enforced by the runtime rather than developer discipline.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

Log in to join the discussion and vote.

Log in

Related stories

0
ProgrammingDEV Community ·

Why Your Next.js ISR Pages Stay Stale for Days Despite a 5-Minute Revalidate

Incremental Static Regeneration (ISR) in Next.js does not refresh pages on a fixed timer — it only triggers a rebuild when a user requests a page that has already exceeded its revalidate window. This means low-traffic pages can serve cached content that is days or even weeks old, regardless of how short the revalidate interval is set. A developer running a Next.js site on Vercel discovered detail pages with a 300-second revalidate were showing cached versions over six days old, simply because few users visited them. Response headers such as x-vercel-cache and age can quickly reveal whether a page is stale and how long ago it was last cached. The behavior is by design and works well for high-traffic pages, but can be problematic for content like prices or availability where accuracy is critical.

0
ProgrammingDEV Community ·

Google Rolls Out September 2026 Spam Update Globally Across All Languages

Google launched its September 2026 spam update on September 24, 2026, at 9:15 AM Pacific Time, with the rollout expected to take up to two weeks to complete. The update applies globally and covers all languages, according to Google's Search Status Dashboard. This marks what independent observers have called Google's fourth spam update of 2026, following earlier releases in June and August. Google has not disclosed which specific spam signals, systems, or content types are targeted by this update. Website owners are advised to monitor ranking changes carefully during the rollout period rather than drawing early conclusions from short-term volatility.

0
ProgrammingDEV Community ·

React Hooks Complete Guide Part 2: Performance, React 19, and Next.js Hooks Explained

Developer Abrar Galib has published Part 2 of a comprehensive React Hooks guide on DEV Community, building on the core and advanced hooks covered in Part 1. This installment covers performance hooks such as useTransition and useDeferredValue, which help keep apps responsive during heavy updates by letting React prioritize urgent work. The guide also details hooks for external data, library authors, and form handling, including useActionState, useOptimistic, and the use hook introduced in React 19. Next.js-specific hooks like useRouter, usePathname, and useLinkStatus are explained alongside newer React 19.3 features such as ViewTransition and Fragment Refs. The guide concludes with a section on custom hooks, common mistakes, and a developer cheatsheet, with all content confirmed compatible with React 19 and its latest stable release.

0
ProgrammingDEV Community ·

How Data Modelling and Schema Design Shape Reliable Power BI Reports

Data modelling in Power BI involves organising tables, keys, and relationships so that business questions can be answered accurately before any visuals or DAX measures are built. A flat-table design, as illustrated by a Kenya Crops Dataset, stores all attributes and facts in a single wide table, which is simple but becomes inefficient as data grows. The star schema, considered the preferred approach for most Power BI solutions, places a central fact table surrounded by dimension tables, balancing simplicity, performance, and maintainability. A snowflake schema further normalises dimensions into additional related tables, which is useful for shared or hierarchical data but should not be over-applied. Choosing the right schema depends on dataset size, complexity, and how long the reporting solution is expected to remain in use.