{
  "slug": "sdf",
  "term": "Software-defined flash (SDF)",
  "aliases": [
    "SDF",
    "Software-defined flash",
    "Baidu SDF"
  ],
  "definition": "SDF is a flash architecture that exposes individual channels and erase operations to host software while retaining per-channel translation and media management.",
  "note": "Interface: The ASPLOS 2014 design uses 8 MB aligned writes and erases with 8 KB reads. Software chooses channels and reclaims blocks; channel engines retain block mapping, dynamic wear leveling, bad-block management and BCH ECC. [paper]\n\nMemory: Removing the device write cache does not mean removing all DRAM: the data path still buffers transfers. Acknowledgment after programming does not establish all-or-nothing recovery for an interrupted write. [paper] [pku]\n\nCapacity: The reported 99% is device-level raw-flash availability. Protection also relies on replication across racks, whose capacity cost is outside that percentage. Eliminating device garbage collection does not eliminate host compaction. [paper]\n\nWorkload: High aggregate throughput needs concurrent channel access. With one slice, batch size one and 512 KB random reads, SDF trails the conventional comparison device. Lower write-latency variation also need not mean lower mean latency. [paper]",
  "status": "unknown",
  "advantages": [],
  "limitations": [],
  "related": [
    "open-channel-ssd",
    "lightnvm",
    "ftl"
  ],
  "supersedes": [],
  "supersededBy": [],
  "firstSeen": null,
  "lastSeen": null,
  "updated": "2026-11-17",
  "citations": [
    {
      "id": "paper",
      "title": "Ouyang et al. SDF: Software-Defined Flash for Web-Scale Internet Storage Systems. ASPLOS 2014, pp. 471-484, sections 2-3.",
      "url": "https://ranger.uta.edu/~sjiang/pubs/papers/ouyang14-SDF.pdf",
      "accessed": "2026-11-17"
    },
    {
      "id": "pku",
      "title": "Peking University copy of SDF, section 2.2: different wording about write completion.",
      "url": "https://ceca.pku.edu.cn/media/lw/65d709d4b0f26917fb3e1640e05ab952.pdf",
      "accessed": "2026-11-17"
    },
    {
      "id": "doi",
      "title": "ACM ASPLOS 2014 proceedings record: SDF, DOI 10.1145/2541940.2541959.",
      "url": "https://doi.org/10.1145/2541940.2541959",
      "accessed": "2026-11-17"
    }
  ],
  "provenance": [
    "Source synthesis of Baidu's 2014 design and workloads, not a hardware reproduction or current-device benchmark. Status is unknown because present availability was not established.",
    "Section 2.2 differs between copies: the UTA-hosted PDF says synchronous operations; the PKU-hosted PDF says atomic operations. Neither wording alone supplies an interrupted-write recovery protocol. [paper] [pku]",
    "The microbenchmark compares 44 synchronous SDF threads against one asynchronous issuing thread for the other devices, with different I/O paths. This is a system comparison, not a controlled FTL-only ablation. [paper]"
  ]
}
