{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,30]],"date-time":"2026-06-30T05:44:39Z","timestamp":1782798279225,"version":"3.54.5"},"reference-count":45,"publisher":"IEEE","license":[{"start":{"date-parts":[[2026,5,18]],"date-time":"2026-05-18T00:00:00Z","timestamp":1779062400000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2026,5,18]],"date-time":"2026-05-18T00:00:00Z","timestamp":1779062400000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100012226","name":"Fundamental Research Funds for the Central Universities","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100012226","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2026,5,18]]},"DOI":"10.1109\/infocom59046.2026.11571521","type":"proceedings-article","created":{"date-parts":[[2026,6,29]],"date-time":"2026-06-29T19:38:15Z","timestamp":1782761895000},"page":"1-10","source":"Crossref","is-referenced-by-count":0,"title":["A Novel Hat-Shaped Device-Cloud Collaborative Inference Framework for Large Language Models"],"prefix":"10.1109","author":[{"given":"Zuan","family":"Xie","sequence":"first","affiliation":[{"name":"University of Science and Technology of China,School of Computer Science and Technology"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yang","family":"Xu","sequence":"additional","affiliation":[{"name":"University of Science and Technology of China,School of Computer Science and Technology"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hongli","family":"Xu","sequence":"additional","affiliation":[{"name":"University of Science and Technology of China,School of Computer Science and Technology"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yunming","family":"Liao","sequence":"additional","affiliation":[{"name":"University of Science and Technology of China,School of Computer Science and Technology"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zhiwei","family":"Yao","sequence":"additional","affiliation":[{"name":"University of Science and Technology of China,School of Computer Science and Technology"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"263","reference":[{"key":"ref1","first-page":"arXiv","article-title":"A survey on effective invocation methods of massive llm services","author":"Wang","year":"2024"},{"key":"ref2","doi-asserted-by":"publisher","DOI":"10.1145\/3649449"},{"key":"ref3","volume-title":"Number of chatgpt users (nov 2023)","author":"Duarte","year":"2023"},{"key":"ref4","first-page":"1049","article-title":"{MArk}: Exploiting cloud services for {Cost-Effective},{SLO-Aware} machine learning inference serving","volume-title":"2019 USENIX Annual Technical Conference (USENIX ATC 19)","author":"Zhang"},{"key":"ref5","article-title":"Personal llm agents: Insights and survey about the capability, efficiency and security","author":"Li","year":"2024"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.1145\/3661167.3661202"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.1109\/TMC.2024.3415661"},{"key":"ref8","article-title":"Apple intelligence foundation language models","author":"Gunter","year":"2024"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.1145\/3662006.3662067"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.1109\/icmlcn64995.2025.11140540"},{"key":"ref11","article-title":"Citer: Collaborative inference for efficient large language model decoding with token-level routing","author":"Zheng","year":"2025"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.1109\/ICMC60390.2024.00008"},{"key":"ref13","article-title":"Split-and-denoise: Protect large language model inference with local differential privacy","author":"Mai","year":"2023"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1109\/MWC.014.2300400"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.48550\/ARXIV.1706.03762"},{"key":"ref16","article-title":"Openai gpt-3: Understanding the architecture"},{"key":"ref17","article-title":"Llama 2: Open foundation and fine-tuned chat models","author":"Touvron","year":"2023"},{"key":"ref18","first-page":"521","article-title":"Orca: A distributed serving system for {Transformer-Based} generative models","volume-title":"16th USENIX Symposium on Operating Systems Design and Implementation (OSDI 22)","author":"Yu"},{"key":"ref19","article-title":"Taming throughput-latency tradeoff in llm inference with sarathi-serve","author":"Agrawal","year":"2024"},{"key":"ref20","article-title":"Deepspeed-fastgen: High-throughput text generation for llms via mii and deepspeed-inference","author":"Holmes","year":"2024"},{"issue":"3","key":"ref21","first-page":"6","article-title":"Vicuna: An open-source chatbot impressing gpt-4 with 90%* chatgpt quality","volume":"2","author":"Chiang","year":"2023"},{"key":"ref22","first-page":"7655","article-title":"Unlocking efficiency in large language model inference: A comprehensive survey of speculative decoding","volume-title":"Findings of the Association for Computational Linguistics ACL 2024","author":"Xia"},{"key":"ref23","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP.2019.8683546"},{"key":"ref24","article-title":"Eagle: Speculative sampling requires rethinking feature uncertainty","author":"Li","year":"2024"},{"key":"ref25","article-title":"A taxonomy of attacks and defenses in split learning","author":"Shabbir","year":"2025"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.14722\/ndss.2025.230030"},{"key":"ref27","doi-asserted-by":"publisher","DOI":"10.1109\/TDSC.2024.3387396"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.1109\/TIFS.2024.3356821"},{"key":"ref29","article-title":"Phantom: Privacy-preserving deep neural network model obfuscation in heterogeneous tee and gpu system","volume-title":"34th USENIX Security Symposium","author":"Bai"},{"key":"ref30","doi-asserted-by":"publisher","DOI":"10.1145\/3638550.3641131"},{"key":"ref31","article-title":"Split ways: Privacy-preserving training of encrypted data using split learning","author":"Khan","year":"2023"},{"key":"ref32","doi-asserted-by":"publisher","DOI":"10.1007\/978-981-97-5606-3_7"},{"key":"ref33","doi-asserted-by":"publisher","DOI":"10.1016\/j.eswa.2024.125494"},{"key":"ref34","doi-asserted-by":"publisher","DOI":"10.1145\/3407023.3407045"},{"issue":"2","key":"ref35","first-page":"2","article-title":"Docker: lightweight linux containers for consistent development and deployment","volume":"239","author":"Merkel","year":"2014","journal-title":"Linux j"},{"key":"ref36","article-title":"Pytorch: An imperative style, high-performance deep learning library","volume":"32","author":"Paszke","year":"2019","journal-title":"Advances in neural information processing systems"},{"key":"ref37","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-540-30218-6_19"},{"key":"ref38","article-title":"Flashattention-2: Faster attention with better parallelism and work partitioning","author":"Dao","year":"2023"},{"key":"ref39","article-title":"Accelerating self-attentions for llm serving with flashinfer","author":"Ye","year":"2024"},{"key":"ref40","article-title":"Nvidia collective communications library (nccl)"},{"key":"ref41","article-title":"Iperf: The tcp\/udp bandwidth measurement tool","author":"Tirumala","year":"1999"},{"key":"ref42","first-page":"1073","article-title":"Get to the point: Summarization with pointer-generator networks","volume-title":"Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics","volume":"1","author":"See"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.1109\/TMC.2023.3331690"},{"key":"ref44","doi-asserted-by":"publisher","DOI":"10.5040\/9781472568083.article-144"},{"key":"ref45","article-title":"Sharegpt","year":"2023"}],"event":{"name":"IEEE INFOCOM 2026 - IEEE Conference on Computer Communications","location":"Tokyo, Japan","start":{"date-parts":[[2026,5,18]]},"end":{"date-parts":[[2026,5,21]]}},"container-title":["IEEE INFOCOM 2026 - IEEE Conference on Computer Communications"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/11571071\/11571169\/11571521.pdf?arnumber=11571521","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,6,30]],"date-time":"2026-06-30T05:12:42Z","timestamp":1782796362000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11571521\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,5,18]]},"references-count":45,"URL":"https:\/\/doi.org\/10.1109\/infocom59046.2026.11571521","relation":{},"subject":[],"published":{"date-parts":[[2026,5,18]]}}}