{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,10,10]],"date-time":"2025-10-10T01:26:25Z","timestamp":1760059585217,"version":"build-2065373602"},"reference-count":37,"publisher":"MDPI AG","issue":"13","license":[{"start":{"date-parts":[[2025,6,24]],"date-time":"2025-06-24T00:00:00Z","timestamp":1750723200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"name":"FCT\u2014Funda\u00e7\u00e3o para a Ci\u00eancia e Tecnologia","award":["C645112083-00000059"],"award-info":[{"award-number":["C645112083-00000059"]}]},{"name":"PRR\u2014Plano de Recupera\u00e7\u00e3o e Resili\u00eancia","award":["C645112083-00000059"],"award-info":[{"award-number":["C645112083-00000059"]}]},{"name":"NextGenerationEU","award":["C645112083-00000059"],"award-info":[{"award-number":["C645112083-00000059"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Electronics"],"abstract":"<jats:p>The adoption of Machine Learning Operations (MLOps) has grown rapidly as organisations seek to streamline the development and deployment of machine learning (ML) models. A core concept in MLOps workflows is the ML pipeline, consisting of a sequence of tasks representing the various stages of the ML lifecycle, such as data preprocessing, model training, and evaluation. As these tasks have different resource requirements and computational demands, using heterogeneous computing environments has become important. However, to exploit this heterogeneity, it is essential to map each task within a pipeline to the right machine. This paper introduces a modular and flexible placement system for ML pipelines that automatically allocates tasks to the most suitable machines in order to reduce execution and waiting times. Although designed to support custom placement strategies, the system employs a two-phase strategy: pipeline scheduling and task placement. During the scheduling phase, the Shortest Job First (SJF) algorithm determines the execution order of the pipelines. In the task placement phase, a heuristic-based method is used to assign tasks to machines. Experimental evaluations across a range of ML models and datasets demonstrate that the proposed system significantly outperforms baseline methods and the Kubernetes default scheduler. It achieved reductions of up to 68% in total execution time, and over 80% in average waiting time. Moreover, the system also demonstrates efficient pipeline dispatching in scenarios where multiple pipelines are submitted for execution. These results highlight the system\u2019s potential to improve resource utilisation and accelerate ML model development in heterogeneous environments.<\/jats:p>","DOI":"10.3390\/electronics14132555","type":"journal-article","created":{"date-parts":[[2025,6,24]],"date-time":"2025-06-24T08:50:57Z","timestamp":1750755057000},"page":"2555","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Optimising ML Pipeline Execution via Smart Task Placement"],"prefix":"10.3390","volume":"14","author":[{"ORCID":"https:\/\/orcid.org\/0009-0009-5916-4062","authenticated-orcid":false,"given":"Pedro","family":"Rodrigues","sequence":"first","affiliation":[{"name":"Instituto de Telecomunica\u00e7\u00f5es, 3810-193 Aveiro, Portugal"}]},{"ORCID":"https:\/\/orcid.org\/0009-0001-7898-7883","authenticated-orcid":false,"given":"Julio","family":"Corona","sequence":"additional","affiliation":[{"name":"Instituto de Telecomunica\u00e7\u00f5es, 3810-193 Aveiro, Portugal"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6504-9441","authenticated-orcid":false,"given":"M\u00e1rio","family":"Antunes","sequence":"additional","affiliation":[{"name":"Instituto de Telecomunica\u00e7\u00f5es, 3810-193 Aveiro, Portugal"},{"name":"Departamento de Eletr\u00f3nica, Telecomunica\u00e7\u00f5es e Inform\u00e1tica, University of Aveiro, 3810-193 Aveiro, Portugal"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-0107-6253","authenticated-orcid":false,"given":"Rui L.","family":"Aguiar","sequence":"additional","affiliation":[{"name":"Instituto de Telecomunica\u00e7\u00f5es, 3810-193 Aveiro, Portugal"},{"name":"Departamento de Eletr\u00f3nica, Telecomunica\u00e7\u00f5es e Inform\u00e1tica, University of Aveiro, 3810-193 Aveiro, Portugal"}]}],"member":"1968","published-online":{"date-parts":[[2025,6,24]]},"reference":[{"key":"ref_1","doi-asserted-by":"crossref","first-page":"115","DOI":"10.1038\/nature21056","article-title":"Dermatologist-level classification of skin cancer with deep neural networks","volume":"542","author":"Esteva","year":"2017","journal-title":"Nature"},{"key":"ref_2","unstructured":"Bojarski, M., Del Testa, D., Dworakowski, D., Firner, B., Flepp, B., Goyal, P., Jackel, L.D., Monfort, M., Muller, U., and Zhang, J. (2016). End to end learning for self-driving cars. arXiv."},{"key":"ref_3","doi-asserted-by":"crossref","first-page":"631","DOI":"10.1016\/j.procs.2020.01.057","article-title":"Credit Card Fraud Detection using Machine Learning Algorithms","volume":"165","author":"Dornadula","year":"2019","journal-title":"Procedia Comput. Sci."},{"key":"ref_4","doi-asserted-by":"crossref","unstructured":"Covington, P., Adams, J., and Sargin, E. (2016, January 15\u201319). Deep Neural Networks for YouTube Recommendations. Proceedings of the 10th ACM Conference on Recommender Systems, New York, NY, USA.","DOI":"10.1145\/2959100.2959190"},{"key":"ref_5","doi-asserted-by":"crossref","first-page":"31866","DOI":"10.1109\/ACCESS.2023.3262138","article-title":"Machine Learning Operations (MLOps): Overview, Definition, and Architecture","volume":"11","author":"Kreuzberger","year":"2023","journal-title":"IEEE Access"},{"key":"ref_6","unstructured":"Eken, B., Pallewatta, S., Tran, N.K., Tosun, A., and Babar, M.A. (2024). A Multivocal Review of MLOps Practices, Challenges and Open Issues. arXiv."},{"key":"ref_7","doi-asserted-by":"crossref","first-page":"e5606","DOI":"10.1002\/cpe.5606","article-title":"RALB-HC: A resource-aware load balancer for heterogeneous cluster","volume":"33","author":"Ahmed","year":"2019","journal-title":"Concurr. Comput. Pract. Exp."},{"key":"ref_8","doi-asserted-by":"crossref","unstructured":"Xin, D., Miao, H., Parameswaran, A., and Polyzotis, N. (2021, January 20\u201325). Production Machine Learning Pipelines: Empirical Analysis and Optimization Opportunities. Proceedings of the 2021 International Conference on Management of Data, New York, NY, USA.","DOI":"10.1145\/3448016.3457566"},{"key":"ref_9","doi-asserted-by":"crossref","first-page":"828","DOI":"10.1007\/s42979-023-02282-2","article-title":"MLOps Challenges in Industry 4.0","volume":"4","author":"Faubel","year":"2023","journal-title":"SN Comput. Sci."},{"key":"ref_10","doi-asserted-by":"crossref","unstructured":"Le, T.N., Sun, X., Chowdhury, M., and Liu, Z. (2020, January 27\u201330). AlloX: Compute allocation in hybrid clusters. Proceedings of the Fifteenth European Conference on Computer Systems, New York, NY, USA.","DOI":"10.1145\/3342195.3387547"},{"key":"ref_11","doi-asserted-by":"crossref","unstructured":"Jayaram Subramanya, S., Arfeen, D., Lin, S., Qiao, A., Jia, Z., and Ganger, G.R. (2023, January 23\u201326). Sia: Heterogeneity-aware, goodput-optimized ML-cluster scheduling. Proceedings of the 29th Symposium on Operating Systems Principles, New York, NY, USA.","DOI":"10.1145\/3600006.3613175"},{"key":"ref_12","doi-asserted-by":"crossref","unstructured":"Chen, H.M., Chen, S.Y., Hsueh, S.H., and Wang, S.K. (July, January 30). Designing an Improved ML Task Scheduling Mechanism on Kubernetes. Proceedings of the 2023 Sixth International Symposium on Computer, Consumer and Control (IS3C), Taichung, Taiwan.","DOI":"10.1109\/IS3C57901.2023.00024"},{"key":"ref_13","doi-asserted-by":"crossref","first-page":"153432","DOI":"10.1109\/ACCESS.2019.2948150","article-title":"SCARL: Attentive Reinforcement Learning-Based Scheduling in a Multi-Resource Heterogeneous Cluster","volume":"7","author":"Cheong","year":"2019","journal-title":"IEEE Access"},{"key":"ref_14","doi-asserted-by":"crossref","first-page":"292","DOI":"10.1016\/j.jpdc.2017.05.001","article-title":"New scheduling approach using reinforcement learning for heterogeneous distributed systems","volume":"117","author":"Orhean","year":"2018","journal-title":"J. Parallel Distrib. Comput."},{"key":"ref_15","unstructured":"Narayanan, D., Santhanam, K., Kazhamiaka, F., Phanishayee, A., and Zaharia, M. (2020, January 4\u20136). Heterogeneity-Aware Cluster Scheduling Policies for Deep Learning Workloads. Proceedings of the 14th USENIX Symposium on Operating Systems Design and Implementation (OSDI 20), Online."},{"key":"ref_16","doi-asserted-by":"crossref","unstructured":"Zhang, K., and Wu, B. (2012, January 24\u201328). Task Scheduling for GPU Heterogeneous Cluster. Proceedings of the 2012 IEEE International Conference on Cluster Computing Workshops, Beijing, China.","DOI":"10.1109\/ClusterW.2012.20"},{"key":"ref_17","doi-asserted-by":"crossref","unstructured":"Symeonidis, G., Nerantzis, E., Kazakis, A., and Papakostas, G.A. (2022, January 26\u201329). MLOps - Definitions, Tools and Challenges. Proceedings of the 2022 IEEE 12th Annual Computing and Communication Workshop and Conference (CCWC), Las Vegas, NV, USA.","DOI":"10.1109\/CCWC54503.2022.9720902"},{"key":"ref_18","doi-asserted-by":"crossref","first-page":"384","DOI":"10.1016\/S0022-0000(75)80008-0","article-title":"NP-complete scheduling problems","volume":"10","author":"Ullman","year":"1975","journal-title":"J. Comput. Syst. Sci."},{"key":"ref_19","doi-asserted-by":"crossref","unstructured":"Kumar, R., Pattnaik, P.K., and Tavares, J.M.R.S. (2023). Computational Complexity and Analysis of Supervised Machine Learning Algorithms. Next Generation of Internet of Things, Springer.","DOI":"10.1007\/978-981-19-1412-6"},{"key":"ref_20","doi-asserted-by":"crossref","unstructured":"Witten, I.H., Frank, E., and Hall, M.A. (2011). Chapter 6\u2014Implementations: Real Machine Learning Schemes. Data Mining: Practical Machine Learning Tools and Techniques, Morgan Kaufmann. [3rd ed.].","DOI":"10.1016\/B978-0-12-374856-0.00006-7"},{"key":"ref_21","first-page":"301","article-title":"Support vector machine solvers","volume":"3","author":"Bottou","year":"2007","journal-title":"Large Scale Kernel Mach."},{"key":"ref_22","unstructured":"Douwes, C., and Serizel, R. (2024). From Computation to Consumption: Exploring the Compute-Energy Link for Training and Testing Neural Networks for SED Systems. arXiv."},{"key":"ref_23","unstructured":"Mei, X., Brei, N., and Lawrence, D. (2023, January 8\u201312). Towards High-Performance AI4NP Applications on Modern GPU Platforms. Proceedings of the 2023 International Conference on Computing in High Energy and Nuclear Physics (CHEP 2023), Norfolk, VA, USA."},{"key":"ref_24","doi-asserted-by":"crossref","unstructured":"Sevilla, J., Heim, L., Ho, A., Besiroglu, T., Hobbhahn, M., and Villalobos, P. (2022, January 18\u201323). Compute Trends Across Three Eras of Machine Learning. Proceedings of the 2022 International Joint Conference on Neural Networks (IJCNN), Padua, Italy.","DOI":"10.1109\/IJCNN55064.2022.9891914"},{"key":"ref_25","unstructured":"Sevilla, J., Heim, L., Hobbhahn, M., Besiroglu, T., Ho, A., and Villalobos, P. (2025, May 04). Estimating Training Compute of Deep Learning Models. Available online: https:\/\/epoch.ai\/blog\/estimating-training-compute."},{"key":"ref_26","unstructured":"Hobbhahn, M., and Sevilla, J. (2025, May 04). What Is the Backward-Forward FLOP Ratio for Neural Networks?. Available online: https:\/\/epoch.ai\/blog\/backward-forward-FLOP-ratio."},{"key":"ref_27","unstructured":"Getzner, J., Charpentier, B., and G\u00fcnnemann, S. (2023). Accuracy is not the only metric that matters: Estimating the energy consumption of deep learning models. arXiv."},{"key":"ref_28","unstructured":"Borji, A. (2021). Enhancing sensor resolution improves CNN accuracy given the same number of parameters or FLOPS. arXiv."},{"key":"ref_29","first-page":"2093","article-title":"Adult","volume":"5","author":"Becker","year":"1996","journal-title":"UCI Mach. Learn. Repos."},{"key":"ref_30","first-page":"547","article-title":"Wine Quality","volume":"47","author":"Cortez","year":"2009","journal-title":"UCI Mach. Learn. Repos."},{"key":"ref_31","first-page":"2473","article-title":"Default of Credit Card Clients","volume":"36","author":"Yeh","year":"2009","journal-title":"UCI Mach. Learn. Repos."},{"key":"ref_32","unstructured":"Stolfo, S., Fan, W., Lee, W., Prodromidis, A., and Chan, P. (1999). KDD Cup 1999 Data. UCI Mach. Learn. Repos."},{"key":"ref_33","doi-asserted-by":"crossref","unstructured":"Moustafa, N., and Slay, J. (2015, January 10\u201312). UNSW-NB15: A comprehensive data set for network intrusion detection systems (UNSW-NB15 network data set). Proceedings of the 2015 Military Communications and Information Systems Conference (MilCIS), Canberra, ACT, Australia.","DOI":"10.1109\/MilCIS.2015.7348942"},{"key":"ref_34","doi-asserted-by":"crossref","first-page":"141","DOI":"10.1109\/MSP.2012.2211477","article-title":"The mnist database of handwritten digit images for machine learning research","volume":"29","author":"Deng","year":"2012","journal-title":"IEEE Signal Process. Mag."},{"key":"ref_35","unstructured":"Xiao, H., Rasul, K., and Vollgraf, R. (2017). Fashion-mnist: A novel image dataset for benchmarking machine learning algorithms. arXiv."},{"key":"ref_36","unstructured":"Krizhevsky, A. (2025, June 21). Learning Multiple Layers of Features from Tiny Images. Available online: https:\/\/www.cs.toronto.edu\/~kriz\/learning-features-2009-TR.pdf."},{"key":"ref_37","doi-asserted-by":"crossref","first-page":"104340","DOI":"10.1016\/j.dib.2019.104340","article-title":"A citrus fruits and leaves dataset for detection and classification of citrus diseases through machine learning","volume":"26","author":"Rauf","year":"2019","journal-title":"Data Brief"}],"container-title":["Electronics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/2079-9292\/14\/13\/2555\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,10,9]],"date-time":"2025-10-09T17:57:44Z","timestamp":1760032664000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/2079-9292\/14\/13\/2555"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,6,24]]},"references-count":37,"journal-issue":{"issue":"13","published-online":{"date-parts":[[2025,7]]}},"alternative-id":["electronics14132555"],"URL":"https:\/\/doi.org\/10.3390\/electronics14132555","relation":{},"ISSN":["2079-9292"],"issn-type":[{"type":"electronic","value":"2079-9292"}],"subject":[],"published":{"date-parts":[[2025,6,24]]}}}