Efficient Fault Tolerance for Recommendation Model Training via Erasure Coding

Summary: ECRec applies erasure coding tailored to DLRM large, sparse embedding tables with a hybrid erasure-code/replication strategy that correctly and efficiently updates redundant parameters. Implemented on XDL, it avoids training pauses on failure, cuts overhead up to 66%, speeds recovery up to 9.8×, and continues with only 7–13% throughput loss. (summarized by gpt-5-mini on Feb 09 2026)

Paper ID: 13339
Venue: VLDB
Year: 2023
Pagerank: 5.7303405e-05
Overall Rank: 6,958 | 52.27%
DOI: 10.14778/3611479.3611514

Incoming Non-self Citations Over Time

Authors

1. Tianyu Zhang (Carnegie Mellon University)
2. Kaige Liu (Carnegie Mellon University)
3. Jack Kosaian (Carnegie Mellon University)
4. Juncheng Yang (Carnegie Mellon University)
5. Rashmi Vinayak (Carnegie Mellon University)

BibTeX Citation

@article{zhang_vldb23,
        title = {{Efficient Fault Tolerance for Recommendation Model Training via Erasure Coding}},
        author = {Zhang, Tianyu and Liu, Kaige and Kosaian, Jack and Yang, Juncheng and Vinayak, Rashmi},
        journal = {PVLDB},
        series = {{VLDB} '23},
        volume = {16},
        number = {11},
        pages = {3137--3150},
        doi = {10.14778/3611479.3611514},
        url = {https://doi.org/10.14778/3611479.3611514},
        year = {2023}
}

Incoming Citations (Sorted by Pagerank)

Showing 1 of 1 citing papers.

Rank	Citing Paper	Year	Venue	Pagerank
10,804	IncrCP: Decomposing and Orchestrating Incremental Checkpoints for Effective Recommendation Model Training	2025	VLDB	5.093636e-05

Outgoing Citations (Sorted by Pagerank)

Showing 3 of 3 cited papers.

Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.

Rank	Cited Paper	Year	Venue	Pagerank
133	A Case for Redundant Arrays of Inexpensive Disks (RAID)	1988	SIGMOD	0.00030370315
2,688	Accelerating Recommendation System Training by Leveraging Popular Choices	2022	VLDB	8.2564305e-05
4,003	XORing Elephants: Novel Erasure Codes for Big Data	2013	VLDB	6.9656006e-05

Semantically Similar Papers

#	Overall Rank	Paper	Year	Venue
1	489	Distributed Representations of Tuples for Entity Resolution	2018	VLDB
2	5,588	Federated Matrix Factorization with Privacy Guarantee	2022	VLDB
3	9,556	CAFE: Towards Compact, Adaptive, and Fast Embedding for Large-scale Recommendation Models	2024	SIGMOD
4	9,520	Experimental Analysis of Large-scale Learnable Vector Storage Compression	2024	VLDB
5	8,907	Scheduling Data Processing Pipelines for Incremental Training on MLP-based Recommendation Models	2025	SIGMOD
6	6,954	DLRover-RM: Resource Optimization for Deep Recommendation Models Training in the Cloud	2024	VLDB
7	2,688	Accelerating Recommendation System Training by Leveraging Popular Choices	2022	VLDB
8	13,327	DECK: Experiences on Delta Checkpointing for Industrial Recommendation Systems	2025	VLDB
9	10,804	IncrCP: Decomposing and Orchestrating Incremental Checkpoints for Effective Recommendation Model Training	2025	VLDB
10	9,164	FEC: Efficient Deep Recommendation Model Training with Flexible Embedding Communication	2023	SIGMOD