Fast Database Restarts at Facebook
Summary: Shared memory decouples memory state from process lifetime in Scuba, enabling fast, weekly upgrades while preserving in-memory data. Recovery is 2–3 minutes per server with no memory-footprint increase, boosting availability and making the approach applicable to other in-memory databases. (summarized by gpt-5-nano on Feb 09 2026)
Incoming Non-self Citations Over Time
Authors
- 1. Aakash Goel (Georgia Institute of Technology; Meta)
- 2. Bhuwan Chopra (Meta)
- 3. Ciprian Gerea (Meta)
- 4. Dhruv Matani (Meta)
- 5. Josh Metzler (Meta)
- 6. Fahim Ul Haq (Meta)
- 7. Janet L. Wiener (Meta)
BibTeX Citation
@inproceedings{goel_sigmod14,
title = {{Fast Database Restarts at Facebook}},
author = {Goel, Aakash and Chopra, Bhuwan and Gerea, Ciprian and Matani, Dhruv and Metzler, Josh and Haq, Fahim Ul and Wiener, Janet L.},
series = {{SIGMOD} '14},
booktitle = {Proceedings of the {ACM} {SIGMOD} International Conference on Management of Data},
publisher = {Association for Computing Machinery},
doi = {10.1145/2588555.2595642},
url = {https://dl.acm.org/doi/10.1145/2588555.2595642},
year = {2014}
}Incoming Citations (Sorted by Pagerank)
Showing 6 of 6 citing papers.
| Rank | Citing Paper | Year | Venue | Pagerank |
|---|---|---|---|---|
| 1,310 | Realtime Data Processing at Facebook | 2016 | SIGMOD | 0.00011065568 |
| 4,534 | Persistent Memory Hash Indexes: An Experimental Evaluation | 2021 | VLDB | 6.5408712e-05 |
| 9,051 | BPF-DB: A Kernel-Embedded Transactional Database Management System For eBPF Applications | 2025 | SIGMOD | 5.2152045e-05 |
| 9,363 | Meta's Next-generation Realtime Monitoring and Analytics Platform | 2022 | VLDB | 5.1727202e-05 |
| 11,704 | The Case for DBMS Live Patching | 2024 | VLDB | 4.9605343e-05 |
| 11,911 | Live Patching Database Management Systems | 2022 | SIGMOD | 4.9605343e-05 |
Previous
Page 1 / 1
Next
Outgoing Citations (Sorted by Pagerank)
Showing 6 of 6 cited papers.
Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.
| Rank | Cited Paper | Year | Venue | Pagerank |
|---|---|---|---|---|
| 12 | C-Store: A Column-oriented DBMS | 2005 | VLDB | 0.00068985264 |
| 49 | Dremel: Interactive Analysis of Web-Scale Datasets | 2010 | VLDB | 0.00043085722 |
| 177 | The Vertica Analytic Database: C-Store 7 Years Later | 2012 | VLDB | 0.00026571747 |
| 574 | Efficient Transaction Processing in SAP HANA Database – The End of a Column Store Myth | 2012 | SIGMOD | 0.00016142754 |
| 1,212 | Processing a Trillion Cells per Mouse Click | 2012 | VLDB | 0.00011493478 |
| 1,379 | Scuba: Diving into Data at Facebook | 2013 | VLDB | 0.00010836332 |
Previous
Page 1 / 1
Next
Semantically Similar Papers
| # | Overall Rank | Paper | Year | Venue |
|---|---|---|---|---|
| 1 | 9,361 | Databases in the Era of Memory-Centric Computing | 2025 | CIDR |
| 2 | 5,528 | A Cost-Effective Method for Providing Improved Data Availability During DBMS Restart Recovery After a Failure | 1993 | VLDB |
| 3 | 4,044 | The Case for Distributed Shared-Memory Databases with RDMA-Enabled Memory Disaggregation | 2023 | VLDB |
| 4 | 1,263 | Fast Checkpoint Recovery Algorithms for Frequently Consistent Applications | 2011 | SIGMOD |
| 5 | 2,303 | In-Memory Performance for Big Data | 2015 | VLDB |
| 6 | 2,642 | Robustness to Crash in a Distributed Database: A Non Shared-Memory Multi-Processor Approach | 1984 | VLDB |
| 7 | 2,155 | A Recovery Algorithm for A High-Performance Memory-Resident Database System | 1987 | SIGMOD |
| 8 | 3,010 | Instant Recovery for Main-Memory Databases | 2015 | CIDR |
| 9 | 10,115 | Recovery Protocols for Shared Memory Database Systems | 1995 | SIGMOD |
| 10 | 1,585 | On the Design and Scalability of Distributed Shared-Data Databases | 2015 | SIGMOD |