backpressure · shipping · also back pressure

Backpressure

Backpressure propagates downstream capacity constraints upstream so a producer slows or pauses further work.

Related Observatory article: Read on hesela.com

Queue boundary: SEDA separates stages with explicit event queues; blocking when a downstream queue is full can propagate congestion upstream. Dropping work is a different response. Buffering can absorb a burst but cannot create downstream service capacity. [1]

Write contract: A healthy Node.js Writable may accept a chunk while write returns false; pause further writes until drain rather than resubmitting that chunk. highWaterMark is a flow-control threshold, not a strict memory cap. Errors still need separate handling. [2]

Persistence: A buffer-pressure signal is not a durable-storage acknowledgement. FileHandle.sync requests a flush to the storage device; its implementation depends on the operating system and device. [3]

Relations

related
Admission control, Queue depth, Coordinated omission

Citations

  1. Welsh, Culler and Brewer: SEDA, SOSP 2001, section 3.3 (finite event queues and overload responses). (accessed 2026-11-20)
  2. Node.js stream documentation: buffering, writable.write and drain. (accessed 2026-11-20)
  3. Node.js file-system documentation: FileHandle.sync and writeFile flush option. (accessed 2026-11-20)
Sources & caveats
  • SEDA is a service-architecture paper, not an SSD benchmark. Node.js API documentation was checked on 2026-11-20.
  • The companion example exercises an in-memory Node.js Writable; it neither writes a storage device nor simulates power loss.

JSON: /data/terms/backpressure.json

Published by Hesela team · Updated

Report a correction · Editorial standards · Reuse guidance

Cite this page
Hesela team. Backpressure. Hesela. 2026-11-20. https://hesela.dev/terms/backpressure/