{
  "slug": "annual-sector-error-rate",
  "term": "Annual sector error rate",
  "aliases": [
    "ASER"
  ],
  "definition": "Annual sector error rate (ASER) normalizes the mean count of latent sector errors per drive in a one-year window by the number of sectors per drive.",
  "note": "Denominator: ASER = E[X(t) - X(t - 12)] / S, with t in months and S sectors per drive; zero-error drives remain in the cohort mean [lse2007].\n\nA constructed example has 100 errors across 1,000 drives, each with 1,000,000 sectors for one year: ASER is 0.0000001 whether one drive or 100 drives contain those errors. The affected-drive fractions are instead 0.1% and 10% [worksheet].\n\nDo not convert capacity-normalized ASER to errors per bit read by multiplying the sector capacity by eight. Read exposure and error-counting conventions are different inputs. An ECC comparison, for example, defines UBER through codeword failure rate divided by codeword length [cai2017].",
  "related": [
    "latent-sector-error",
    "uncorrectable-bit-error-rate",
    "data-scrubbing",
    "afr"
  ],
  "citations": [
    {
      "id": "lse2007",
      "title": "Bairavasundaram et al. An Analysis of Latent Sector Errors in Disk Drives. SIGMETRICS 2007, section 5.3.1 and Figure 2.",
      "url": "https://research.cs.wisc.edu/adsl/Publications/latent-sigmetrics07.pdf",
      "accessed": "2026-11-06"
    },
    {
      "id": "cai2017",
      "title": "Cai et al. Error Characterization, Mitigation, and Recovery in Flash Memory Based Solid-State Drives (2017), section VI.C.",
      "url": "https://research.ece.cmu.edu/safari/pubs/1706.08642.pdf",
      "accessed": "2026-11-06"
    },
    {
      "id": "worksheet",
      "title": "Hesela: ASER denominator examples with complete synthetic one-year cohorts.",
      "url": "https://hesela.com/analyses/scrubbing-detection-not-durability/#aser-denominators",
      "accessed": "2026-11-06"
    }
  ],
  "status": "research",
  "advantages": [],
  "limitations": [],
  "supersedes": [],
  "supersededBy": [],
  "firstSeen": null,
  "lastSeen": null,
  "updated": "2026-11-06",
  "provenance": [
    "Figure 2 uses a cohort observed for 24 months, reporting its first and second years separately [lse2007].",
    "The example is arithmetic on invented complete annual counts, not a fleet estimate or a replication of the private traces [worksheet]."
  ]
}