{
  "slug": "delayed-hit",
  "term": "Delayed cache hit",
  "aliases": [
    "Delayed hit"
  ],
  "definition": "A delayed cache hit is a request for a nonresident object that joins an outstanding fetch and waits for its completion rather than initiating another fetch.",
  "note": "Not fetching again does not mean that data are ready. Separate resident hits, delayed hits and fetch-triggering misses when interpreting a cache metric. [delayed-paper]\n\nHesela's constructed four-read example has one 10 ms fetch in every row. Arrivals at 0, 1, 2 and 3 ms produce waits of 10, 9, 8 and 7 ms, averaging 8.5 ms. Spacing arrivals to 0, 10, 20 and 30 ms instead gives 2.5 ms. Completion precedes equal-time arrivals; service after the fetch is instantaneous. These are model values, not device measurements. [hesela]",
  "related": [
    "request-coalescing",
    "sieve-eviction",
    "miss-ratio-curve"
  ],
  "citations": [
    {
      "id": "delayed-paper",
      "title": "Atre et al. Caching with Delayed Hits. SIGCOMM 2020, sections 2 and 6. DOI 10.1145/3387514.3405883.",
      "url": "https://www.cs.cmu.edu/~weinaw/pdf/delayed-hits.pdf",
      "accessed": "2026-10-19"
    },
    {
      "id": "hesela",
      "title": "Hesela: single-object delayed-fill arithmetic, event table and explicit assumptions.",
      "url": "https://hesela.com/analyses/sieve-cache-eviction-traces/#delayed-hits",
      "accessed": "2026-10-20"
    }
  ],
  "status": "research",
  "advantages": [],
  "limitations": [],
  "supersedes": [],
  "supersededBy": [],
  "firstSeen": null,
  "lastSeen": null,
  "updated": "2026-10-20"
}