Beluga: A CXL-Based Memory Architecture for Scalable and Efficient LLM KVCache Management

Summary: Beluga exploits CXL switches to expose a shared, large-scale memory pool with native load/store access for GPU/CPU KVCache, avoiding RDMA’s latency/protocol overhead. Beluga-KVCache uses this architecture to scale long-context LLM inference, cutting TTFT 89.6% and boosting vLLM throughput 7.35x. (summarized by gpt-5-mini on Apr 11 2026)

Paper ID: 7453
Venue: SIGMOD
Year: 2026
Pagerank: 4.1945683e-05
Overall Rank: 10,143 | 29.44%
DOI: 10.1145/3786627

Incoming Non-self Citations Over Time

No non-self incoming citations found for this paper in this database.

Authors

Incoming Citations (Sorted by Pagerank)

Showing 0 of 0 citing papers.

Rank	Citing Paper	Year	Venue	Pagerank

Outgoing Citations (Sorted by Pagerank)

Showing 17 of 17 cited papers.

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

Rank	Cited Paper	Year	Venue	Pagerank
1,593	PolarFS: An Ultra-low Latency and Failure Resilient Distributed File System for Shared Storage Cloud Database	2018	VLDB	0.00011224049
1,872	ReAcTable: Enhancing ReAct for Table Question Answering	2024	VLDB	0.00010259702
2,572	Efficient Distributed Memory Management with RDMA and Caching	2018	VLDB	8.519943e-05
3,859	OpenSearch-SQL: Enhancing Text-to-SQL with Dynamic Few-shot and Consistency Alignment	2025	SIGMOD	6.6907933e-05
3,936	Rethinking Database High Availability with RDMA Networks	2019	VLDB	6.6162264e-05
4,544	ScaleStore: A Fast and Cost-Efficient Storage Engine using DRAM, NVMe, and RDMA	2022	SIGMOD	6.1000636e-05
5,042	Design Guidelines for Correct, Efficient, and Scalable Synchronization using One-Sided RDMA	2023	SIGMOD	5.7414429e-05
6,223	Distributed GPU Joins on Fast RDMA-capable Networks	2023	SIGMOD	5.1496398e-05
6,741	DEX: Scalable Range Indexing on Disaggregated Memory	2024	VLDB	4.9432931e-05
6,796	InferDB: In-Database Machine Learning Inference Using Indexes	2024	VLDB	4.9241624e-05
7,061	Serving Deep Learning Models with Deduplication from Relational Databases	2022	VLDB	4.8463881e-05
7,339	SpareLLM: Automatically Selecting Task-Specific Minimum-Cost Large Language Models under Equivalence Constraint	2025	SIGMOD	4.7579469e-05
8,001	Rethinking Stateful Stream Processing with RDMA	2022	SIGMOD	4.6092573e-05
8,649	Zero-sided RDMA: Network-driven Data Shuffling for Disaggregated Heterogeneous Cloud DBMSs	2024	SIGMOD	4.4762914e-05
8,950	Unlocking the Potential of CXL for Disaggregated Memory in Cloud-Native Databases	2025	SIGMOD	4.4231907e-05
9,476	Adda: Towards Efficient in-Database Feature Generation via LLM-based Agents	2025	SIGMOD	4.3341665e-05
10,782	From Scale-Up to Scale-Out: PolarDB's Journey to Achieving 2 Billion tpmC	2025	VLDB	4.1945683e-05

Semantically Similar Papers

Overall Rank	Paper	Year	Venue	Pagerank
10,222	RetroInfer: A Vector Storage Engine for Scalable Long-Context LLM Inference	2026	VLDB	4.1945683e-05
10,066	DepCache: A KV Cache Management Framework for GraphRAG with Dependency Attention	2026	SIGMOD	4.1945683e-05
13,138	Database Perspective on LLM Inference Systems	2025	VLDB	-
8,513	CXL Memory Performance for In-Memory Data Processing	2025	VLDB	4.4947795e-05
13,088	OrbitFlow: SLO-Aware Long-Context LLM Serving with Fine-Grained KV Cache Reconfiguration	2026	VLDB	-
3,565	Cache-Craft: Managing Chunk-Caches for Efficient Retrieval-Augmented Generation	2025	SIGMOD	6.9655362e-05
9,805	MEMO: Fine-grained Tensor Management For Ultra-long Context LLM Training	2025	SIGMOD	4.2805224e-05
10,020	HotPrefix: Hotness-Aware KV Cache Scheduling for Efficient Prefix Sharing in LLM Inference Systems	2026	SIGMOD	4.1945683e-05
8,950	Unlocking the Potential of CXL for Disaggregated Memory in Cloud-Native Databases	2025	SIGMOD	4.4231907e-05
6,357	PQCache: Product Quantization-based KVCache for Long Context LLM Inference	2025	SIGMOD	5.0970739e-05