EPIC: Abstraction and Polymorphism of In-Network Collectives on Ethernet
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866918509802422272 |
|---|---|
| author | Yuan, Yitao Nie, Jianglong Bai, Tianyu Zhou, Ruizhe Cao, Siyuan Fan, Xujie Xu, Yuchen Chen, Junkai Zhao, Chenqi Zhang, Nengyuan Fang, Shaoke Chen, Jiangyuan Chen, Yuanfeng Sun, Jiaqi Wang, Zhan Xu, Xiaohua Zhang, Yuchao Liu, Yang Yang, Xiangrui Lin, Jing Hu, Xiaohe Li, Yang Jiang, Chao Xiao, Limin Zhang, Weifeng Wang, Junjie Cheng, Wei Lan, Yazhu Dong, Jianbo Fu, Binzhang Wu, Wenfei |
| author_facet | Yuan, Yitao Nie, Jianglong Bai, Tianyu Zhou, Ruizhe Cao, Siyuan Fan, Xujie Xu, Yuchen Chen, Junkai Zhao, Chenqi Zhang, Nengyuan Fang, Shaoke Chen, Jiangyuan Chen, Yuanfeng Sun, Jiaqi Wang, Zhan Xu, Xiaohua Zhang, Yuchao Liu, Yang Yang, Xiangrui Lin, Jing Hu, Xiaohe Li, Yang Jiang, Chao Xiao, Limin Zhang, Weifeng Wang, Junjie Cheng, Wei Lan, Yazhu Dong, Jianbo Fu, Binzhang Wu, Wenfei |
| contents | In-Network Collective (INC) acceleration holds immense potential for optimizing AI training and inference; however, its cross-layer nature has historically hindered investment and adoption within the open Ethernet ecosystem. To bridge this gap, we propose EPIC (Ethernet Polymorphic In-network Collective), an INC protocol specification and reference system built on the principle of "Unified Abstraction, Polymorphic Realization." EPIC introduces an abstraction compatible with standard Ethernet that aligns functional boundaries with participant roles, while offering polymorphic realizations tailored to varying hardware capabilities.
We address three fundamental challenges: first, we employ a modular design that enables an evolutionary path from simple to complex implementations, allowing vendors to iterate their hardware incrementally; second, we apply formal verification methodologies to prove the correctness of all proposed polymorphic modes; and third, we develop a unified resource management model versatile enough for diverse INC scenarios. Extensive validation -- spanning model checking, packet/flow simulations, VM emulation, Tofino Testbed, and FPGA/RTL verification -- confirms EPIC's correctness, performance gain, and feasibility. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_18683 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | EPIC: Abstraction and Polymorphism of In-Network Collectives on Ethernet Yuan, Yitao Nie, Jianglong Bai, Tianyu Zhou, Ruizhe Cao, Siyuan Fan, Xujie Xu, Yuchen Chen, Junkai Zhao, Chenqi Zhang, Nengyuan Fang, Shaoke Chen, Jiangyuan Chen, Yuanfeng Sun, Jiaqi Wang, Zhan Xu, Xiaohua Zhang, Yuchao Liu, Yang Yang, Xiangrui Lin, Jing Hu, Xiaohe Li, Yang Jiang, Chao Xiao, Limin Zhang, Weifeng Wang, Junjie Cheng, Wei Lan, Yazhu Dong, Jianbo Fu, Binzhang Wu, Wenfei Distributed, Parallel, and Cluster Computing In-Network Collective (INC) acceleration holds immense potential for optimizing AI training and inference; however, its cross-layer nature has historically hindered investment and adoption within the open Ethernet ecosystem. To bridge this gap, we propose EPIC (Ethernet Polymorphic In-network Collective), an INC protocol specification and reference system built on the principle of "Unified Abstraction, Polymorphic Realization." EPIC introduces an abstraction compatible with standard Ethernet that aligns functional boundaries with participant roles, while offering polymorphic realizations tailored to varying hardware capabilities. We address three fundamental challenges: first, we employ a modular design that enables an evolutionary path from simple to complex implementations, allowing vendors to iterate their hardware incrementally; second, we apply formal verification methodologies to prove the correctness of all proposed polymorphic modes; and third, we develop a unified resource management model versatile enough for diverse INC scenarios. Extensive validation -- spanning model checking, packet/flow simulations, VM emulation, Tofino Testbed, and FPGA/RTL verification -- confirms EPIC's correctness, performance gain, and feasibility. |
| title | EPIC: Abstraction and Polymorphism of In-Network Collectives on Ethernet |
| topic | Distributed, Parallel, and Cluster Computing |
| url | https://arxiv.org/abs/2605.18683 |