The DOI matches its Crossref record.
Fast crash recovery in RAMCloud
Abstract
RAMCloud is a DRAM-based storage system that provides inexpensive durability and availability by recovering quickly after crashes, rather than storing replicas in DRAM. RAMCloud scatters backup data across hundreds or thousands of disks, and it harnesses hundreds of servers in parallel to reconstruct lost data. The system uses a log-structured approach for all its data, in DRAM as well as on disk: this provides high performance both during normal operation and during recovery. RAMCloud employs randomized techniques to manage the system in a scalable and decentralized fashion. In a 60-node cluster, RAMCloud recovers 35 GB of data from a failed server in 1.6 seconds. Our measurements suggest that the approach will scale to recover larger memory sizes (64 GB or more) in less time with larger clusters.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“We consider p99 latency since it is important to control the tail latency in LLM serving as with all other user-facing services DeCandia et al. (2007); Ongaro et al. (2011).”
From Preble: Efficient Distributed Prompt Scheduling for LLM Serving (2024) · cites this paper as Ongaro et al. (2011)
Citations
Cites
No references could be resolved yet.
Research works printed but not linked: 21
- 10.1145/1629087.1629088 · DOI
- 10.1145/195058.195412 · DOI
- 10.1145/1365815.1365816 · DOI
- 10.14778/1454159.1454167 · DOI
- 10.1145/1807128.1807130 · DOI
- 10.1145/1294261.1294281 · DOI
- 10.1145/602259.602261 · DOI
- 10.1109/69.180602 · DOI
- 10.1145/945445.945450 · DOI
- 10.1145/78969.78972 · DOI
- Hunt P., 2010, Proceedings of the 2010 USENIX annual technical conference, USENIX ATC '10, 11
- Johnson R., 2009, Personal Communications, March, 24, 20
- 10.14778/1454159.1454211 · DOI
- Mitzenmacher M. D., 1996, University of California
- 10.1145/1965724.1965751 · DOI
- 10.1109/2.16 · DOI
- 10.1145/50202.50214 · DOI
- 10.1145/146941.146943 · DOI
- Seltzer M., 1995, Proceedings of the USENIX 1995 Technical Conference, TCON'95, 21
- 10.1109/MSST.2010.5496972 · DOI
- 10.1109/TNET.2002.808407 · DOI
Cited by
Sorted by citations
Preble: Efficient Distributed Prompt Scheduling for LLM Serving
Srivatsa, V., He, Z., Abhyankar, R., Li, D., & Zhang, Y. (2024) · arXiv · doi:10.48550/arxiv.2407.00023 · Cited by 1
“We consider p99 latency since as all other user-facing services [7, 33], it is important to prevent high tail latency.”
Background
Metadata
- Authors
- Ongaro, D., Rumble, S. M., Stutsman, R., Ousterhout, J., & Rosenblum, M.
- Affiliations
- Stanford University (US)
- ORCID
- Stutsman, Ryan
- Venue
- Proceedings of the Twenty-Third ACM Symposium on Operating Systems Principles
- Publisher
- ACM
- Published
- 2011-10-23
- Pages
- 29-41
- OpenAlex
- W2003597767
- Cited by
- 360
- References
- 0
- Citations by year
- 2026: 1 · 2025: 7 · 2024: 5 · 2023: 10 · 2022: 12 · 2021: 19 · 2020: 23 · 2019: 24 · 2018: 38 · 2017: 50 · 2016: 40 · 2015: 44 · 2014: 26 · 2013: 35 · 2012: 26
- FWCI
- 24.78
- Type
- conference paper
- Language
- en
- Open access
- closed
- Locations
- ACM SIGOPS Symposium on Operating Systems Principles (SOSP) (publishedVersion)
- Funders
- Division of Computer and Network Systems (CNS-0716806CNS-1052985)
- Topics
- Physical Sciences › Computer Science › Computer Networks and Communications › Advanced Data Storage Technologies
- Physical Sciences › Computer Science › Computer Networks and Communications › Distributed systems and fault tolerance
- Physical Sciences › Computer Science › Information Systems › Advanced Database Systems and Queries
- Keywords
- crash recovery, log-structured storage, data reconstruction, distributed storage, data durability, availability, scalability, randomized algorithms, decentralized management
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
