{"project":[{"grant_number":"761493","name":"5G Development and validation platform for global industry-specific network services and Apps","_id":"28"},{"_id":"1","name":"SFB 901"},{"_id":"4","name":"SFB 901 - Project Area C"},{"name":"SFB 901 - Subproject C4","_id":"16"}],"citation":{"short":"S.B. Schneider, N.P. Satheeschandran, M. Peuster, H. Karl, in: IEEE Conference on Network Softwarization (NetSoft), IEEE, 2020.","ieee":"S. B. Schneider, N. P. Satheeschandran, M. Peuster, and H. Karl, “Machine Learning for Dynamic Resource Allocation in Network Function Virtualization,” in <i>IEEE Conference on Network Softwarization (NetSoft)</i>, Ghent, Belgium, 2020.","bibtex":"@inproceedings{Schneider_Satheeschandran_Peuster_Karl_2020, title={Machine Learning for Dynamic Resource Allocation in Network Function Virtualization}, booktitle={IEEE Conference on Network Softwarization (NetSoft)}, publisher={IEEE}, author={Schneider, Stefan Balthasar and Satheeschandran, Narayanan Puthenpurayil and Peuster, Manuel and Karl, Holger}, year={2020} }","ama":"Schneider SB, Satheeschandran NP, Peuster M, Karl H. Machine Learning for Dynamic Resource Allocation in Network Function Virtualization. In: <i>IEEE Conference on Network Softwarization (NetSoft)</i>. IEEE; 2020.","chicago":"Schneider, Stefan Balthasar, Narayanan Puthenpurayil Satheeschandran, Manuel Peuster, and Holger Karl. “Machine Learning for Dynamic Resource Allocation in Network Function Virtualization.” In <i>IEEE Conference on Network Softwarization (NetSoft)</i>. IEEE, 2020.","mla":"Schneider, Stefan Balthasar, et al. “Machine Learning for Dynamic Resource Allocation in Network Function Virtualization.” <i>IEEE Conference on Network Softwarization (NetSoft)</i>, IEEE, 2020.","apa":"Schneider, S. B., Satheeschandran, N. P., Peuster, M., &#38; Karl, H. (2020). Machine Learning for Dynamic Resource Allocation in Network Function Virtualization. In <i>IEEE Conference on Network Softwarization (NetSoft)</i>. Ghent, Belgium: IEEE."},"has_accepted_license":"1","ddc":["000"],"publisher":"IEEE","oa":"1","type":"conference","department":[{"_id":"75"}],"_id":"16219","file_date_updated":"2020-03-03T11:42:16Z","status":"public","date_created":"2020-03-03T11:42:22Z","year":"2020","author":[{"orcid":"0000-0001-8210-4011","id":"35343","first_name":"Stefan Balthasar","full_name":"Schneider, Stefan Balthasar","last_name":"Schneider"},{"first_name":"Narayanan Puthenpurayil","last_name":"Satheeschandran","full_name":"Satheeschandran, Narayanan Puthenpurayil"},{"first_name":"Manuel","id":"13271","last_name":"Peuster","full_name":"Peuster, Manuel"},{"id":"126","first_name":"Holger","last_name":"Karl","full_name":"Karl, Holger"}],"user_id":"35343","title":"Machine Learning for Dynamic Resource Allocation in Network Function Virtualization","publication":"IEEE Conference on Network Softwarization (NetSoft)","abstract":[{"lang":"eng","text":"Network function virtualization (NFV) proposes\r\nto replace physical middleboxes with more flexible virtual\r\nnetwork functions (VNFs). To dynamically adjust to everchanging\r\ntraffic demands, VNFs have to be instantiated and\r\ntheir allocated resources have to be adjusted on demand.\r\nDeciding the amount of allocated resources is non-trivial.\r\nExisting optimization approaches often assume fixed resource\r\nrequirements for each VNF instance. However, this can easily\r\nlead to either waste of resources or bad service quality if too\r\nmany or too few resources are allocated.\r\n\r\nTo solve this problem, we train machine learning models\r\non real VNF data, containing measurements of performance\r\nand resource requirements. For each VNF, the trained models\r\ncan then accurately predict the required resources to handle\r\na certain traffic load. We integrate these machine learning\r\nmodels into an algorithm for joint VNF scaling and placement\r\nand evaluate their impact on resulting VNF placements. Our\r\nevaluation based on real-world data shows that using suitable\r\nmachine learning models effectively avoids over- and underallocation\r\nof resources, leading to up to 12 times lower resource\r\nconsumption and better service quality with up to 4.5 times\r\nlower total delay than using standard fixed resource allocation."}],"conference":{"name":"IEEE Conference on Network Softwarization (NetSoft)","location":"Ghent, Belgium"},"file":[{"file_id":"16220","content_type":"application/pdf","access_level":"open_access","date_created":"2020-03-03T11:42:16Z","date_updated":"2020-03-03T11:42:16Z","file_name":"ris_preprint.pdf","relation":"main_file","file_size":476590,"creator":"stschn"}],"language":[{"iso":"eng"}],"date_updated":"2022-01-06T06:52:46Z"}