Efficient LLM Serving for Agentic Workflows: A Data Systems Perspective
Summary: Helium recasts agentic LLM workflows as query plans with LLM calls as operators, enabling cross-call reuse of prompts, KV states, and intermediate results. Proactive caching and cache-aware scheduling yield up to 1.56× speedup over existing agent-serving systems. (summarized by gpt-5.6-luna on Jul 26 2026)
Incoming Non-self Citations Over Time
No non-self incoming citations found for this paper in this database.
Authors
- 1. Noppanat Wadlom (National University of Singapore)
- 2. Junyi Shen (National University of Singapore)
- 3. Yao Lu (National University of Singapore)
BibTeX Citation
@inproceedings{wadlom_sigmod26,
title = {{Efficient LLM Serving for Agentic Workflows: A Data Systems Perspective}},
author = {Wadlom, Noppanat and Shen, Junyi and Lu, Yao},
series = {{SIGMOD} '26},
booktitle = {Proceedings of the {ACM} {SIGMOD} International Conference on Management of Data},
publisher = {Association for Computing Machinery},
doi = {10.1145/3802046},
url = {https://dl.acm.org/doi/10.1145/3802046},
year = {2026}
}Incoming Citations (Sorted by Pagerank)
Showing 0 of 0 citing papers.
| Rank | Citing Paper | Year | Venue | Pagerank |
|---|
Previous
Page 1 / 1
Next
Outgoing Citations (Sorted by Pagerank)
Showing 1 of 1 cited papers.
Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.
| Rank | Cited Paper | Year | Venue | Pagerank |
|---|---|---|---|---|
| 6,210 | Supporting Our AI Overlords: Redesigning Data Systems to be Agent-First | 2026 | CIDR | 5.9425753e-05 |
Previous
Page 1 / 1
Next
Semantically Similar Papers
| # | Overall Rank | Paper | Year | Venue |
|---|---|---|---|---|
| 1 | 13,291 | OrbitFlow: SLO-Aware Long-Context LLM Serving with Fine-Grained KV Cache Reconfiguration | 2026 | VLDB |
| 2 | 13,375 | Reimagining Deep Learning Systems Through the Lens of Data Systems | 2024 | VLDB |
| 3 | 13,343 | Database Perspective on LLM Inference Systems | 2025 | VLDB |
| 4 | 7,439 | AOP: Automated and Interactive LLM Pipeline Orchestration for Answering Complex Queries | 2025 | CIDR |
| 5 | 10,733 | ScaleLLM: A Technique for Scalable LLM-augmented Data Systems | 2025 | SIGMOD |
| 6 | 10,191 | AlignedServe: Orchestrating Prefix-aware Batching to Build a High-throughput and Computing-efficient LLM Serving System | 2026 | SIGMOD |
| 7 | 6,101 | LLM for Data Management | 2024 | VLDB |
| 8 | 13,288 | Serving Hybrid LLM Loads with SLO Guarantees Using CPU-GPU Attention Piggybacking | 2026 | SIGMOD |
| 9 | 6,210 | Supporting Our AI Overlords: Redesigning Data Systems to be Agent-First | 2026 | CIDR |
| 10 | 6,900 | Apt-Serve: Adaptive Request Scheduling on Hybrid Cache for Scalable LLM Inference Serving | 2025 | SIGMOD |