{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,20]],"date-time":"2026-09-20T17:21:09Z","timestamp":1789924869557,"version":"4.0.1"},"reference-count":64,"publisher":"Elsevier BV","license":[{"start":{"date-parts":[[2027,1,1]],"date-time":"2027-01-01T00:00:00Z","timestamp":1798761600000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/www.elsevier.com\/tdm\/userlicense\/1.0\/"},{"start":{"date-parts":[[2027,1,1]],"date-time":"2027-01-01T00:00:00Z","timestamp":1798761600000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/www.elsevier.com\/legal\/tdmrep-license"},{"start":{"date-parts":[[2026,7,21]],"date-time":"2026-07-21T00:00:00Z","timestamp":1784592000000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/100018693","name":"Horizon Europe","doi-asserted-by":"publisher","award":["101015922"],"award-info":[{"award-number":["101015922"]}],"id":[{"id":"10.13039\/100018693","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100010661","name":"Horizon 2020 Framework Programme","doi-asserted-by":"publisher","id":[{"id":"10.13039\/100010661","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100013209","name":"Hellenic Foundation for Research and Innovation","doi-asserted-by":"publisher","award":["5349"],"award-info":[{"award-number":["5349"]}],"id":[{"id":"10.13039\/501100013209","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["elsevier.com","sciencedirect.com"],"crossmark-restriction":true},"short-container-title":["Future Generation Computer Systems"],"published-print":{"date-parts":[[2027,1]]},"DOI":"10.1016\/j.future.2026.108708","type":"journal-article","created":{"date-parts":[[2026,7,14]],"date-time":"2026-07-14T15:17:14Z","timestamp":1784042234000},"page":"108708","update-policy":"https:\/\/doi.org\/10.1016\/elsevier_cm_policy","source":"Crossref","is-referenced-by-count":0,"special_numbering":"C","title":["CECAIServe: Facilitating ML inference serving across the cloud\u2013edge-continuum"],"prefix":"10.1016","volume":"186","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-1347-980X","authenticated-orcid":false,"given":"Aimilios","family":"Leftheriotis","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Achilleas","family":"Tzenetopoulos","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"George","family":"Lentaris","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Dimitrios","family":"Soudris","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"George","family":"Theodoridis","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"78","reference":[{"key":"10.1016\/j.future.2026.108708_b1","doi-asserted-by":"crossref","DOI":"10.1016\/j.future.2024.107544","article-title":"Connected vehicles ecological driving based on deep reinforce learning: Application of web 3.0 technologies in traffic optimization","volume":"163","author":"Ma","year":"2025","journal-title":"Future Gener. Comput. Syst."},{"key":"10.1016\/j.future.2026.108708_b2","doi-asserted-by":"crossref","DOI":"10.1016\/j.future.2025.107821","article-title":"Towards sustainable smart cities: Workflow scheduling in cloud of health things (coht) using deep reinforcement learning and moth flame optimization for edge\u2013cloud systems","volume":"170","author":"Khaleel","year":"2025","journal-title":"Future Gener. Comput. Syst."},{"key":"10.1016\/j.future.2026.108708_b3","doi-asserted-by":"crossref","DOI":"10.1016\/j.future.2025.107874","article-title":"Adaptive heuristics for scheduling dnn inferencing on edge and cloud for personalized uav fleets","volume":"173","author":"Raj","year":"2025","journal-title":"Future Gener. Comput. Syst."},{"key":"10.1016\/j.future.2026.108708_b4","unstructured":"A. Robustelli, A. Marfoglia, C. D\u2019Errico, S. Mellone, A. Carbonaro, et al., Feasibility of mlops-based healthcare pipelines in ensuring the cybersecurity framework, in: CEUR WORKSHOP PROCEEDINGS, Ceur Workshop proceedings, 2025, pp. 1\u201312."},{"key":"10.1016\/j.future.2026.108708_b5","series-title":"2023 ACM\/IEEE International Symposium on Empirical Software Engineering and Measurement","first-page":"1","article-title":"Assessing the use of automl for data-driven software engineering","author":"Calefato","year":"2023"},{"key":"10.1016\/j.future.2026.108708_b6","unstructured":"C. Daoud, D. Azar, J. Boich\u00e9, C. Urtado, S. Vauttier, Skeltymlops: Orchestrating collaborative mlops activities, in: MLOps25-Workshop on Machine Learning Operations\/the 1st ECAI Workshop on MLOps, 2025."},{"key":"10.1016\/j.future.2026.108708_b7","series-title":"Special issue on integration of machine learning and edge computing for next generation of smart wearable systems","author":"Gastaldo","year":"2025"},{"key":"10.1016\/j.future.2026.108708_b8","unstructured":"Tensorflow serving, https:\/\/www.tensorflow.org\/tfx\/guide\/serving."},{"key":"10.1016\/j.future.2026.108708_b9","unstructured":"Nvidia triton inference server, https:\/\/developer.nvidia.com\/triton-inference-server."},{"key":"10.1016\/j.future.2026.108708_b10","unstructured":"Torchserve, https:\/\/pytorch.org\/serve\/index.html."},{"key":"10.1016\/j.future.2026.108708_b11","unstructured":"Amazon sagemaker, https:\/\/aws.amazon.com\/sagemaker\/."},{"key":"10.1016\/j.future.2026.108708_b12","unstructured":"Kserve, https:\/\/kserve.github.io\/website\/latest\/."},{"key":"10.1016\/j.future.2026.108708_b13","unstructured":"Bentoml, https:\/\/www.bentoml.com\/."},{"key":"10.1016\/j.future.2026.108708_b14","unstructured":"Litserve, https:\/\/github.com\/Lightning-AI\/LitServe."},{"key":"10.1016\/j.future.2026.108708_b15","unstructured":"Rayserve, https:\/\/docs.ray.io\/en\/latest\/serve\/index.html."},{"key":"10.1016\/j.future.2026.108708_b16","unstructured":"Tensorflow lite, https:\/\/www.tensorflow.org\/lite."},{"key":"10.1016\/j.future.2026.108708_b17","unstructured":"Onnx runtime with tensorrt backend, https:\/\/onnxruntime.ai\/docs\/execution-providers\/TensorRT-ExecutionProvider.html."},{"key":"10.1016\/j.future.2026.108708_b18","series-title":"2024 Joint European Conference on Networks and Communications & 6G Summit","first-page":"931","article-title":"Tf2aif: Facilitating development and deployment of accelerated ai models on the cloud\u2013edge continuum","author":"Leftheriotis","year":"2024"},{"key":"10.1016\/j.future.2026.108708_b19","unstructured":"Onnx runtime with openvino backend, https:\/\/onnxruntime.ai\/docs\/execution-providers\/OpenVINO-ExecutionProvider.html."},{"key":"10.1016\/j.future.2026.108708_b20","unstructured":"Vitis-ai, https:\/\/www.xilinx.com\/products\/design-tools\/vitis\/vitis-ai.html."},{"key":"10.1016\/j.future.2026.108708_b21","unstructured":"F. Romero, Q. Li, N.J. Yadwadkar, C. Kozyrakis, {INFaaS}: Automated model-less inference serving, in: 2021 USENIX Annual Technical Conference, USENIX ATC 21, 2021, pp. 397\u2013411."},{"key":"10.1016\/j.future.2026.108708_b22","unstructured":"D. Crankshaw, X. Wang, G. Zhou, M.J. Franklin, J.E. Gonzalez, I. Stoica, Clipper: A {Low-Latency} online prediction serving system, in: 14th USENIX Symposium on Networked Systems Design and Implementation, NSDI 17, 2017, pp. 613\u2013627."},{"key":"10.1016\/j.future.2026.108708_b23","series-title":"2021 IEEE\/ACM Symposium on Edge Computing","first-page":"271","article-title":"Deeprt: A soft real time scheduler for computer vision applications on the edge","author":"Yang","year":"2021"},{"key":"10.1016\/j.future.2026.108708_b24","unstructured":"W. Cui, H. Zhao, Q. Chen, H. Wei, Z. Li, D. Zeng, C. Li, M. Guo, {DVABatch}: Diversity-aware {Multi-Entry}{Multi-Exit} batching for efficient processing of {DNN} services on {GPUs}, in: 2022 USENIX Annual Technical Conference, USENIX ATC 22, 2022, pp. 183\u2013198."},{"key":"10.1016\/j.future.2026.108708_b25","unstructured":"C. Zhang, M. Yu, W. Wang, F. Yan, {MArk}: Exploiting cloud services for {Cost-Effective}, {SLO-Aware} machine learning inference serving, in: 2019 USENIX Annual Technical Conference, USENIX ATC 19, 2019, pp. 1049\u20131062."},{"key":"10.1016\/j.future.2026.108708_b26","unstructured":"A. Gujarati, R. Karimi, S. Alzayat, W. Hao, A. Kaufmann, Y. Vigfusson, J. Mace, Serving {DNNs} like clockwork: Performance predictability from the bottom up, in: 14th USENIX Symposium on Operating Systems Design and Implementation, OSDI 20, 2020, pp. 443\u2013462."},{"key":"10.1016\/j.future.2026.108708_b27","series-title":"Comparison and benchmarking of ai models and frameworks on mobile devices","author":"Luo","year":"2020"},{"key":"10.1016\/j.future.2026.108708_b28","unstructured":"Ina3221 power monitor, https:\/\/www.ti.com\/lit\/ds\/symlink\/ina3221.pdf?ts=1736147131164."},{"key":"10.1016\/j.future.2026.108708_b29","unstructured":"Amd\/xilinx board utility tool (xbutil), https:\/\/xilinx.github.io\/XRT\/master\/html\/xbutil.html."},{"key":"10.1016\/j.future.2026.108708_b30","unstructured":"Nvidia management library, (nvml), https:\/\/developer.nvidia.com\/management-library-nvml."},{"key":"10.1016\/j.future.2026.108708_b31","series-title":"2012 41st International Conference on Parallel Processing Workshops","first-page":"262","article-title":"Measuring energy and power with papi","author":"Weaver","year":"2012"},{"key":"10.1016\/j.future.2026.108708_b32","unstructured":"Ina226 power monitor, https:\/\/www.ti.com\/lit\/ds\/symlink\/ina226.pdf."},{"key":"10.1016\/j.future.2026.108708_b33","series-title":"Design Patterns: Elements of Reusable Object-Oriented Software","author":"Gamma","year":"1995"},{"key":"10.1016\/j.future.2026.108708_b34","doi-asserted-by":"crossref","DOI":"10.1109\/ACCESS.2026.3673012","article-title":"Navigating the edge-cloud continuum: A state-of-practice survey","author":"Belcastro","year":"2026","journal-title":"IEEE Access"},{"key":"10.1016\/j.future.2026.108708_b35","series-title":"Squeezenet: Alexnet-level accuracy with 50x fewer parameters and < 0.5 mb model size","author":"Iandola","year":"2016"},{"key":"10.1016\/j.future.2026.108708_b36","doi-asserted-by":"crossref","unstructured":"A. Howard, M. Sandler, G. Chu, L.-C. Chen, B. Chen, M. Tan, W. Wang, Y. Zhu, R. Pang, V. Vasudevan, et al., Searching for mobilenetv3, in: Proceedings of the IEEE\/CVF International Conference on Computer Vision, 2019, pp. 1314\u20131324.","DOI":"10.1109\/ICCV.2019.00140"},{"key":"10.1016\/j.future.2026.108708_b37","series-title":"Mobilenets: Efficient convolutional neural networks for mobile vision applications","author":"Howard","year":"2017"},{"key":"10.1016\/j.future.2026.108708_b38","series-title":"International Conference on Machine Learning","first-page":"6105","article-title":"Efficientnet: Rethinking model scaling for convolutional neural networks","author":"Tan","year":"2019"},{"key":"10.1016\/j.future.2026.108708_b39","doi-asserted-by":"crossref","unstructured":"G. Huang, Z. Liu, L. Van Der Maaten, K.Q. Weinberger, Densely connected convolutional networks, in: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2017, pp. 4700\u20134708.","DOI":"10.1109\/CVPR.2017.243"},{"key":"10.1016\/j.future.2026.108708_b40","doi-asserted-by":"crossref","unstructured":"K. He, X. Zhang, S. Ren, J. Sun, Deep residual learning for image recognition, in: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016, pp. 770\u2013778.","DOI":"10.1109\/CVPR.2016.90"},{"key":"10.1016\/j.future.2026.108708_b41","doi-asserted-by":"crossref","unstructured":"C. Szegedy, V. Vanhoucke, S. Ioffe, J. Shlens, Z. Wojna, Rethinking the inception architecture for computer vision, in: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016, pp. 2818\u20132826.","DOI":"10.1109\/CVPR.2016.308"},{"key":"10.1016\/j.future.2026.108708_b42","doi-asserted-by":"crossref","unstructured":"Z. Liu, H. Mao, C.-Y. Wu, C. Feichtenhofer, T. Darrell, S. Xie, A convnet for the 2020s, in: Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition, 2022, pp. 11976\u201311986.","DOI":"10.1109\/CVPR52688.2022.01167"},{"key":"10.1016\/j.future.2026.108708_b43","doi-asserted-by":"crossref","unstructured":"J. Long, E. Shelhamer, T. Darrell, Fully convolutional networks for semantic segmentation, in: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2015, pp. 3431\u20133440.","DOI":"10.1109\/CVPR.2015.7298965"},{"key":"10.1016\/j.future.2026.108708_b44","series-title":"Yolov5 by ultralytics","author":"Jocher","year":"2020"},{"key":"10.1016\/j.future.2026.108708_b45","doi-asserted-by":"crossref","unstructured":"A.H. Lang, S. Vora, H. Caesar, L. Zhou, J. Yang, O. Beijbom, Pointpillars: Fast encoders for object detection from point clouds, in: Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition, 2019, pp. 12697\u201312705.","DOI":"10.1109\/CVPR.2019.01298"},{"key":"10.1016\/j.future.2026.108708_b46","series-title":"Wav2letter: an end-to-end convnet-based speech recognition system","author":"Collobert","year":"2016"},{"key":"10.1016\/j.future.2026.108708_b47","unstructured":"Tensorrt-llm, https:\/\/docs.nvidia.com\/tensorrt-llm\/index.html."},{"key":"10.1016\/j.future.2026.108708_b48","unstructured":"Nvidia tensor cores, https:\/\/www.nvidia.com\/en-eu\/data-center\/tensor-cores\/."},{"issue":"5","key":"10.1016\/j.future.2026.108708_b49","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1145\/3527155","article-title":"Split computing and early exiting for deep learning applications: Survey and research challenges","volume":"55","author":"Matsubara","year":"2022","journal-title":"ACM Comput. Surv."},{"key":"10.1016\/j.future.2026.108708_b50","series-title":"ICASSP 2023-2023 IEEE International Conference on Acoustics, Speech and Signal Processing","first-page":"1","article-title":"Benchmarking convolutional neural network inference on low-power edge devices","author":"Ferraz","year":"2023"},{"key":"10.1016\/j.future.2026.108708_b51","doi-asserted-by":"crossref","unstructured":"A. Abouelhamayed, S. Balle, D. Singh, M. Abdelfattah, Beyond inference: Performance analysis of dnn server overheads for computer vision, in: Proceedings of the 61st ACM\/IEEE Design Automation Conference, 2024, pp. 1\u20136.","DOI":"10.1145\/3649329.3655960"},{"key":"10.1016\/j.future.2026.108708_b52","series-title":"Taming the titans: A survey of efficient llm inference serving","author":"Zhen","year":"2025"},{"key":"10.1016\/j.future.2026.108708_b53","unstructured":"A. Marfoglia, A. Robustelli, C. D\u2019Errico, S. Mellone, A. Carbonaro, et al., A llmops-driven framework for clinical data harmonization, in: CEUR WORKSHOP PROCEEDINGS, Ceur Workshop proceedings, 2025, pp. 25\u201336."},{"key":"10.1016\/j.future.2026.108708_b54","series-title":"2024 9th International Conference on Smart and Sustainable Technologies","first-page":"1","article-title":"Large language model operations (llmops): Definition, challenges, and lifecycle management","author":"Diaz-De-Arcaya","year":"2024"},{"key":"10.1016\/j.future.2026.108708_b55","doi-asserted-by":"crossref","unstructured":"W. Kwon, Z. Li, S. Zhuang, Y. Sheng, L. Zheng, C.H. Yu, J.E. Gonzalez, H. Zhang, I. Stoica, Efficient memory management for large language model serving with pagedattention, in: Proceedings of the ACM SIGOPS 29th Symposium on Operating Systems Principles, 2023.","DOI":"10.1145\/3600006.3613165"},{"key":"10.1016\/j.future.2026.108708_b56","unstructured":"Y. Zhong, S. Liu, J. Chen, J. Hu, Y. Zhu, X. Liu, X. Jin, H. Zhang, {DistServe}: Disaggregating prefill and decoding for goodput-optimized large language model serving, in: 18th USENIX Symposium on Operating Systems Design and Implementation, OSDI 24, 2024, pp. 193\u2013210."},{"key":"10.1016\/j.future.2026.108708_b57","doi-asserted-by":"crossref","first-page":"68658","DOI":"10.52202\/079017-2193","article-title":"Flashattention-3: Fast and accurate attention with asynchrony and low-precision","volume":"37","author":"Shah","year":"2024","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"10.1016\/j.future.2026.108708_b58","unstructured":"llama.cpp, https:\/\/github.com\/ggml-org\/llama.cpp."},{"key":"10.1016\/j.future.2026.108708_b59","series-title":"Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing: System Demonstrations","first-page":"38","article-title":"Transformers: State-of-the-art natural language processing","author":"Wolf","year":"2020"},{"issue":"8","key":"10.1016\/j.future.2026.108708_b60","first-page":"9","article-title":"Language models are unsupervised multitask learners","volume":"1","author":"Radford","year":"2019","journal-title":"OpenAI Blog"},{"key":"10.1016\/j.future.2026.108708_b61","series-title":"Introducing qwen1.5","author":"Team","year":"2024"},{"key":"10.1016\/j.future.2026.108708_b62","series-title":"Gemma 2: Improving open language models at a practical size","author":"Team","year":"2024"},{"key":"10.1016\/j.future.2026.108708_b63","series-title":"Phi-3 technical report: A highly capable language model locally on your phone","author":"Abdin","year":"2024"},{"key":"10.1016\/j.future.2026.108708_b64","doi-asserted-by":"crossref","unstructured":"D. Mendoza, F. Romero, Q. Li, N.J. Yadwadkar, C. Kozyrakis, Interference-aware scheduling for inference serving, in: Proceedings of the 1st Workshop on Machine Learning and Systems, 2021, pp. 80\u201388.","DOI":"10.1145\/3437984.3458837"}],"container-title":["Future Generation Computer Systems"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/api.elsevier.com\/content\/article\/PII:S0167739X26003420?httpAccept=text\/xml","content-type":"text\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/api.elsevier.com\/content\/article\/PII:S0167739X26003420?httpAccept=text\/plain","content-type":"text\/plain","content-version":"vor","intended-application":"text-mining"}],"deposited":{"date-parts":[[2026,9,20]],"date-time":"2026-09-20T16:33:09Z","timestamp":1789921989000},"score":1,"resource":{"primary":{"URL":"https:\/\/linkinghub.elsevier.com\/retrieve\/pii\/S0167739X26003420"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2027,1]]},"references-count":64,"alternative-id":["S0167739X26003420"],"URL":"https:\/\/doi.org\/10.1016\/j.future.2026.108708","relation":{},"ISSN":["0167-739X"],"issn-type":[{"value":"0167-739X","type":"print"}],"subject":[],"published":{"date-parts":[[2027,1]]},"assertion":[{"value":"Elsevier","name":"publisher","label":"This article is maintained by"},{"value":"CECAIServe: Facilitating ML inference serving across the cloud\u2013edge-continuum","name":"articletitle","label":"Article Title"},{"value":"Future Generation Computer Systems","name":"journaltitle","label":"Journal Title"},{"value":"https:\/\/doi.org\/10.1016\/j.future.2026.108708","name":"articlelink","label":"CrossRef DOI link to publisher maintained version"},{"value":"article","name":"content_type","label":"Content Type"},{"value":"\u00a9 2026 The Authors. Published by Elsevier B.V.","name":"copyright","label":"Copyright"}],"article-number":"108708"}}