@techreport{irtf-nmrg-ai-deploy-03,
    number =    {draft-irtf-nmrg-ai-deploy-03},
    type =      {Internet-Draft},
    institution =   {Internet Engineering Task Force},
    publisher = {Internet Engineering Task Force},
    note =      {Work in Progress},
    url =       {https://datatracker.ietf.org/doc/draft-irtf-nmrg-ai-deploy/03/},
    author =    {Yong-Geun Hong and Joo-Sang Youn and Seung-Woo Hong and Pedro Martinez-Julia and Qin Wu},
    title =     {{Considerations of network/system for AI services}},
    pagetotal = 27,
    year =      2026,
    month =     jul,
    day =       6,
    abstract =  {As the development of AI technology has matured and AI technology has begun to be applied in various fields, the execution environment has evolved from dedicated high-performance servers to commodity servers and affordable, small-scale hardware, including microcontrollers, low-performance CPUs, and AI chipsets. This document outlines how to configure the network and system for an AI inference service, providing AI services in a distributed manner. It also outlines the factors to consider when a client connects to a cloud server and an edge device to request an AI service. It describes some use cases for deploying network-based AI services, such as self-driving vehicles and network digital twins.},
}