This repository contains the open-source artifacts for our ACM SIGCOMM 2026 paper, “EPIC: Abstraction and Polymorphism of In-Network Collectives on Ethernet.”
EPIC (Ethernet Polymorphic In-network Collectives) is an in-network collective communication protocol and reference system for the open Ethernet ecosystem. Following the principle of unified abstraction, polymorphic realization, EPIC defines a common abstraction and interoperable interfaces while providing three data-plane modes for devices with different hardware capabilities. It supports six collective primitives: AllReduce, Reduce, Broadcast, AllGather, ReduceScatter, and Barrier.
| Directory | Description |
|---|---|
verify/ |
TLA+/TLC formal verification of EPIC-II and EPIC-III under packet loss, reordering, and duplication. |
flow-sim/ |
OMNeT++ flow-level simulations for evaluating resource management and job completion time. |
pkt-sim/ |
ns-3 packet-level simulations of Ring, EPIC-II, and EPIC-III, including packet-loss experiments. |
emulation/ |
VM-based implementation used to evaluate EPIC's collective modes over RoCE. |
tofino/ |
Tofino prototype, including the host frontend/backend and the P4 switch implementation. |
np/ |
Network-processor implementations of EPIC-II and EPIC-III for AllReduce and Barrier. |
Please refer to the README in each subdirectory for prerequisites, build instructions, and usage.
@inbook{10.1145/3789240.3829196,
author = {Yuan, Yitao and Nie, Jianglong and Bai, Tianyu and Zhou, Ruizhe and Cao, Siyuan and Fan, Xujie and Xu, Yuchen and Chen, Junkai and Zhao, Chenqi and Zhang, Nengyuan and Fang, Shaoke and Chen, Jiangyuan and Chen, Yuanfeng and Sun, Jiaqi and Wang, Zhan and Xu, Xiaohua and Zhang, Yuchao and Liu, Yang and Yang, Xiangrui and Lin, Jing and Hu, Xiaohe and Li, Yang and Jiang, Chao and Xiao, Limin and Zhang, Weifeng and Wang, Junjie and Cheng, Wei and Lan, Yazhu and Dong, Jianbo and Fu, Binzhang and Wu, Wenfei},
title = {EPIC: Abstraction and Polymorphism of In-Network Collectives on Ethernet},
year = {2026},
isbn = {9798400724671},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
url = {https://doi.org/10.1145/3789240.3829196},
abstract = {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 Collectives), 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. We conduct extensive experiments—model checking, packet/flow simulations, VM emulation, Tofino/NP Testbed, and FPGA/RTL verification—to validate EPIC's correctness, performance gains, and feasibility.},
booktitle = {Proceedings of the ACM SIGCOMM 2026 Conference},
pages = {368–395},
numpages = {28}
}