A link-layer error recovery mechanism combining forward error correction with automatic retransmission, in which a receiver retains and combines failed transmission attempts with subsequent retransmissions to improve decoding success beyond what either attempt alone would achieve.

Key Points

Combines two error-handling techniques — forward error correction (adding redundancy that lets some errors be corrected without retransmission) and automatic repeat request (retransmitting when correction isn't sufficient) — into a single hybrid mechanism
Unlike simple ARQ, which discards a failed transmission attempt entirely, HARQ typically retains the failed attempt and combines it with the retransmission, improving the odds of successful decoding beyond what either attempt alone would achieve
Operates within strict timing budgets defined by the underlying radio standard — the round-trip time available for a retransmission to arrive and be processed before it's no longer useful
Propagation delay directly constrains HARQ's usability: a link with round-trip delay exceeding the standard's HARQ timing assumptions requires extended or modified HARQ timers, which is precisely the adaptation 5G NTN architecture had to make for satellite propagation delay
A core reliability mechanism operating largely transparently to applications, functioning as part of the radio link layer rather than something application-level software interacts with directly

Concept

When a receiver fails to correctly decode a transmission even after applying forward error correction, rather than discarding that failed attempt entirely and waiting for a clean retransmission, HARQ retains the failed attempt and combines it with the retransmitted data — since the failed attempt still carries some correct information even if not enough to fully decode — improving the combined odds of successful decoding relative to either transmission considered alone.

Explainer

The timing assumptions built into HARQ are calibrated around terrestrial propagation delays measured in microseconds to low milliseconds, which is precisely why satellite and NTN links — where round-trip propagation delay for GEO can exceed 250 milliseconds — break HARQ's standard timers unless the retransmission window is specifically extended or the mechanism is disabled for those links. This is a concrete, well-documented example of why extending terrestrial cellular standards to non-terrestrial links required more than simply pointing the same radio protocol at a satellite instead of a base station: HARQ timing is one of several link-layer assumptions terrestrial-designed protocols make that satellite propagation delay directly violates, requiring deliberate protocol adaptation rather than transparent reuse. For CE verticals, HARQ's practical relevance is mostly invisible at the application level — it operates as part of the underlying radio link reliability mechanism — but its adaptation, or lack thereof, directly affects the achievable link reliability and effective throughput of any cellular or NTN connection under degraded channel conditions, making it a meaningful factor in overall link performance even though it's rarely a direct configuration point for end users or applications.

Cite this entry

HARQ. ConnectedEarth Knowledge Base. https://connectedearth.space/knowledge-base/harq/ (updated 13 September 2026).