FAST '27 Call for Papers | USENIX

Sponsored by USENIX, the Advanced Computing Systems Association.
The 25th USENIX Conference on File and Storage Technologies (FAST '27) will take place on February 23–25, 2027, at the Hyatt Regency Lake Washington in Renton, WA, USA.
The 25th USENIX Conference on File and Storage Technologies (FAST '27) brings together researchers and practitioners to explore new directions in the design, implementation, evaluation, and deployment of systems related to storage. The program committee interprets storage-related systems broadly: submissions on low-level storage devices, distributed storage systems, information and data management, as well as other systems interconnected with storage are all of interest. The conference will consist of technical presentations including refereed papers and poster sessions.
The topics of interest to FAST are various aspects of systems related to storage, including both core storage topics and the application of storage to different application domains. These include and overlap with, but are not limited to, the following topics.
Core storage topics, such as:
Purpose-built storage systems, such as:
Scalable storage systems, such as:
Emerging storage technologies, such as:
Deployment experience in areas such as:
Authors have the option to submit to one of two submission deadlines. Our intention is to provide authors the flexibility to submit their work in its best form. Papers accepted during either deadline will be presented at the conference and will be published in the proceedings. Pre-publication PDFs accepted in the Spring review cycle will be available following the cycle's conclusion.
Please submit your short and long papers by one of the submission deadlines listed above in PDF format via the submission form. Do not email submissions. There is no separate deadline for abstract submissions.
The program committee and external reviewers will judge papers on technical merit, significance, relevance, and presentation. Research papers on new and unexplored problems are encouraged. A good research paper:
Program committee members, USENIX, and the broader community generally value a paper more highly if it is accompanied by artifacts not previously available. These artifacts may include traces, original data, source code, or tools developed as part of the submitted work.
Blind reviewing of all papers will be done by the program committee, assisted by outside referees when necessary. Accepted papers will be shepherded by a member of the program committee.
In addition to long papers (up to 12 pages), FAST also solicits short papers (up to 6 pages long). Just like long papers, short papers should describe completed research where the problem statement, proposed solution, and evaluation are all logically complete and conclusions are drawn. While short papers are held to the same high standards as long papers, they tend to comprise smaller contributions that require a shorter description and less analysis than in long papers. Papers that describe early-stage research that is not yet fully evaluated are not suitable for this short-paper category. That is, preliminary or work-in-progress papers generally considered in workshops such as HotStorage do not fall within the scope of short papers.
For short papers, the title should be prefixed with "Short Paper: ", followed by the title. The prefix will not be published in the proceedings and short papers will not be called out as such in the program. Authors must also indicate that they are submitting a short paper by checking the appropriate checkbox on the submission form. The program committee will not accept a paper on the condition of adjusting its length beyond typical shepherding guidelines. Submissions will be considered only in the category in which they are submitted.
To see examples of qualified short papers, we encourage authors to browse short papers accepted at prior FAST conferences.
FAST also solicits papers that describe real operational systems, including systems currently in production. Deployed-systems papers should address experience with the practical design, implementation, analysis, deployment, or operation of such systems. We encourage the submission of papers that disprove or strengthen existing assumptions, deepen the understanding of existing problems, and validate known techniques in environments in which they were never before used or tested, with preference given to experimental results based on production data. Deployed-systems papers will be treated similarly to other papers for publication purposes; they need not present new ideas or results to be accepted but should offer useful guidance to practitioners.
A good deployed-systems paper:
For deployed-systems papers, the title should be prefixed with "Deployed System: ", followed by the title. The prefix will not be published in the proceedings. Authors must also indicate that they are submitting a deployed-systems paper by checking the appropriate checkbox on the submission form. If a paper is both short and falls in the deployed-systems category, both prefixes should be used (in any order), and both checkboxes selected.
Double-blind policy for deployed-systems papers: All submissions for FAST '27 are required to follow the double-blind policy (see above). However, with deployed-systems papers, the product or company described in the paper need not be anonymized (unlike author names).
FAST '27 will allow authors to respond to reviews prior to final decision, according to the schedule above. Authors must limit their response to correcting factual errors in the reviews, to addressing questions posed by reviewers, and to clarifying the ideas in the paper. Responses may include new experiments and data in response to a reviewer's request. Responses are optional and limited to 1000 words. FAST will be enforcing a hard limit on the length of the author's response for fairness and to reduce workload (for both authors and reviewers): exceeding the wo