{
  "slug": "fail-slow-drives-software-scheduling",
  "title": "80% of 315 verified fail-slow drives at Alibaba were caused by software scheduling, and 216 of them were in two clusters",
  "kind": "dataset",
  "summary": "Lu et al., FAST 2023 (Perseus), monitored 248K Alibaba drives for 10 months. Of 315 verified fail-slow drives, 252 were caused by software scheduling and 63 by hardware, and 216 of the 252 were in two clusters. Cross-check rows come from Gunawi et al., FAST 2018 (101 reports, 12 institutions).",
  "license": "Small derived table of figures printed in Lu et al. (FAST 2023) and Gunawi et al. (FAST 2018), with credit. Not a Creative Commons license. The USENIX papers remain with their authors or publishers. This file is not a copy of the papers and not the data.",
  "licenseUrl": "https://www.usenix.org/system/files/fast23-lu.pdf",
  "sourceUrl": "https://www.usenix.org/system/files/fast23-lu.pdf",
  "accessed": "2026-10-11",
  "rowCount": 15,
  "columns": [
    {
      "name": "measure",
      "type": "string",
      "description": "The quantity as the source names it."
    },
    {
      "name": "value_low",
      "type": "number",
      "description": "Lower end of a printed range. Empty when the source prints a single value."
    },
    {
      "name": "value_high",
      "type": "number",
      "description": "Single printed value, or the upper end of a range. For 'about' it is the stated figure."
    },
    {
      "name": "unit",
      "type": "string",
      "description": "Unit of the value columns: drives, months, reports, percent, or fraction."
    },
    {
      "name": "scope",
      "type": "string",
      "description": "Population and window the value applies to."
    },
    {
      "name": "note",
      "type": "string",
      "description": "What the value is not, or the source wording behind it."
    },
    {
      "name": "citation_id",
      "type": "string",
      "description": "Id of the opened source in the citations list."
    }
  ],
  "downloadPath": "/data/datasets/fail-slow-drives-software-scheduling.csv",
  "humanPage": "https://hesela.com/analyses/fail-slow-drives-software-scheduling/",
  "limits": [
    "One operator's drives and scheduling software.",
    "216 of the 252 software-caused drives were in two open-channel SSD clusters, so the 80% share is concentrated.",
    "Only 15 of the 63 hardware cases have a vendor root cause.",
    "Precision and recall are on a benchmark built mostly from drives the detector flagged.",
    "The cross-check rows are incident reports from other institutions with other definitions."
  ],
  "citations": [
    {
      "id": "lu-fast23-pdf",
      "title": "Perseus: A Fail-Slow Detection Framework for Cloud Storage Systems, Lu, Xu, Zhang, Zhu, Zhu, Wang, Zhu, Xue, Shu, Li, and Wu, USENIX FAST 2023 (USENIX PDF, sections 5 and 6)",
      "url": "https://www.usenix.org/system/files/fast23-lu.pdf",
      "accessed": "2026-10-11"
    },
    {
      "id": "gunawi-fast18-pdf",
      "title": "Fail-Slow at Scale: Evidence of Hardware Performance Faults in Large Production Systems, Gunawi et al., USENIX FAST 2018 (USENIX PDF, sections 2 to 3)",
      "url": "https://www.usenix.org/system/files/conference/fast18/fast18-gunawi.pdf",
      "accessed": "2026-10-11"
    }
  ]
}