paperbot · PL 论文追踪

全部论文 待读 收藏

P2 Architecting Automotive Software for the Emerging Compute-in-the-Network Paradigm

为新兴的 Compute-in-the-Network 范式构建汽车软件架构
ICSE Proceedings of the 2026 IEEE/ACM 48th International Conference on Software Engineering2026匹配度 2/10 · 论文聚焦汽车网络内计算范式的软件架构建模与领域特定语言,属于汽车系统/架构建模方向,未涉及用户核心关注的静态分析、LLM/Agent 程序分析、浮点程序分析或贝叶斯程序分析,因此匹配度较低。引用 0 inbox
Saad Mubeen, Mohammad Ashjaei, John Lundbäck
crossref: 2026-09-12

AI 速览

一句话:主张将车载网络交换机模型作为一等软件架构实体,以支持网络内计算范式的汽车软件建模。

新兴的车载网络交换机具备计算能力,可将部分应用功能托管并执行于通信结构中,从而降低时延、提升时间关键系统的响应性。然而,现有汽车架构描述语言与组件模型缺乏表示驻留于网络元素中应用软件的抽象,限制了准确的设计与分析。本文论证应将车载网络交换机模型视为一等软件架构实体,并勾勒出若干研究方向:统一抽象、端到端时序模型抽取、与标准及工具链的互操作,以及应对新的网络安全挑战。论文属于愿景/立场论述,未报告实证评估。

问题
现有汽车架构描述语言和组件模型无法表示驻留在网络交换机中的应用软件,导致网络内计算系统难以被准确设计与分析。
方法
通过立场论证,提出将车载网络交换机模型视为一等软件架构实体,并规划统一抽象、端到端时序模型抽取、标准与工具链互操作及网络安全等研究方向。
评估
论文以动机论述和研究议程为主,未给出原型实现或实证评估。
原文摘要(Abstract)

Emerging compute-capable automotive network switches now allow parts of application functionality to be hosted and executed within the communication fabric, complementing the primary application software execution on electronic control units. This compute-in-the-network paradigm offers opportunities for reducing latency and improving responsiveness in time‑critical automotive systems. However, existing automotive architecture description languages and component models lack abstractions for representing application software residing in network elements, limiting accurate design and analysis. This paper motivates the need to treat the models of onboard network switches as the first-class software architectural entities and outlines research opportunities including unified abstractions, end-to-end timing models extraction, alignment with standards and toolchain interoperability, and addressing new cybersecurity challenges.

链接与引用

DOI 原文 · PDF(开放获取) ·

BibTeX
@inproceedings{paperbot11564,
  title = {Architecting Automotive Software for the Emerging Compute-in-the-Network Paradigm},
  author = {Saad Mubeen and Mohammad Ashjaei and John Lundbäck},
  booktitle = {ICSE},
  year = {2026},
  doi = {10.1145/3774748.3795872}
}