[{"title":"Practical Network Conditions for the Convergence of Distributed Optimization","project":[{"name":"SFB 901 - C4: SFB 901 - Subproject C4","_id":"16"},{"name":"SFB 901: SFB 901","_id":"1"},{"_id":"4","name":"SFB 901 - C: SFB 901 - Project Area C"}],"department":[{"_id":"75"}],"date_updated":"2022-11-18T10:05:14Z","language":[{"iso":"eng"}],"ddc":["006"],"user_id":"477","volume":55,"status":"public","has_accepted_license":"1","date_created":"2022-08-16T09:12:55Z","author":[{"last_name":"Redder","id":"52265","first_name":"Adrian","orcid":"https://orcid.org/0000-0001-7391-4688","full_name":"Redder, Adrian"},{"first_name":"Arunselvan","orcid":"https://orcid.org/ 0000-0001-7547-8111","full_name":"Ramaswamy, Arunselvan","last_name":"Ramaswamy","id":"66937"},{"full_name":"Karl, Holger","first_name":"Holger","id":"126","last_name":"Karl"}],"publisher":"Elsevier","publication":"IFAC-PapersOnLine","file_date_updated":"2022-08-31T07:06:30Z","file":[{"file_size":298395,"creator":"aredder","file_id":"33236","content_type":"application/pdf","date_updated":"2022-08-31T07:06:30Z","success":1,"relation":"main_file","date_created":"2022-08-31T07:06:30Z","file_name":"NecSys2022____Practical_Conditions_for_Conv.pdf","access_level":"closed"}],"issue":"13","_id":"32854","intvolume":" 55","conference":{"name":"IFAC Conference on Networked Systems"},"year":"2022","citation":{"short":"A. Redder, A. Ramaswamy, H. Karl, IFAC-PapersOnLine 55 (2022) 133–138.","ieee":"A. Redder, A. Ramaswamy, and H. Karl, “Practical Network Conditions for the Convergence of Distributed Optimization,” IFAC-PapersOnLine, vol. 55, no. 13, pp. 133–138, 2022.","chicago":"Redder, Adrian, Arunselvan Ramaswamy, and Holger Karl. “Practical Network Conditions for the Convergence of Distributed Optimization.” IFAC-PapersOnLine 55, no. 13 (2022): 133–138.","apa":"Redder, A., Ramaswamy, A., & Karl, H. (2022). Practical Network Conditions for the Convergence of Distributed Optimization. IFAC-PapersOnLine, 55(13), 133–138.","ama":"Redder A, Ramaswamy A, Karl H. Practical Network Conditions for the Convergence of Distributed Optimization. IFAC-PapersOnLine. 2022;55(13):133–138.","mla":"Redder, Adrian, et al. “Practical Network Conditions for the Convergence of Distributed Optimization.” IFAC-PapersOnLine, vol. 55, no. 13, Elsevier, 2022, pp. 133–138.","bibtex":"@article{Redder_Ramaswamy_Karl_2022, title={Practical Network Conditions for the Convergence of Distributed Optimization}, volume={55}, number={13}, journal={IFAC-PapersOnLine}, publisher={Elsevier}, author={Redder, Adrian and Ramaswamy, Arunselvan and Karl, Holger}, year={2022}, pages={133–138} }"},"type":"journal_article","page":"133–138"},{"title":"Use What You Know: Network and Service Coordination Beyond Certainty","project":[{"_id":"1","name":"SFB 901: SFB 901"},{"_id":"4","name":"SFB 901 - C: SFB 901 - Project Area C"},{"_id":"16","name":"SFB 901 - C4: SFB 901 - Subproject C4"}],"department":[{"_id":"75"}],"oa":"1","date_updated":"2022-01-11T08:44:04Z","language":[{"iso":"eng"}],"ddc":["004"],"user_id":"35343","abstract":[{"text":"Modern services often comprise several components, such as chained virtual network functions, microservices, or\r\nmachine learning functions. Providing such services requires to decide how often to instantiate each component, where to place these instances in the network, how to chain them and route traffic through them. \r\nTo overcome limitations of conventional, hardwired heuristics, deep reinforcement learning (DRL) approaches for self-learning network and service management have emerged recently. These model-free DRL approaches are more flexible but typically learn tabula rasa, i.e., disregard existing understanding of networks, services, and their coordination. \r\n\r\nInstead, we propose FutureCoord, a novel model-based AI approach that leverages existing understanding of networks and services for more efficient and effective coordination without time-intensive training. FutureCoord combines Monte Carlo Tree Search with a stochastic traffic model. This allows FutureCoord to estimate the impact of future incoming traffic and effectively optimize long-term effects, taking fluctuating demand and Quality of Service (QoS) requirements into account. Our extensive evaluation based on real-world network topologies, services, and traffic traces indicates that FutureCoord clearly outperforms state-of-the-art model-free and model-based approaches with up to 51% higher flow success ratios.","lang":"eng"}],"date_created":"2022-01-11T08:43:26Z","has_accepted_license":"1","status":"public","file_date_updated":"2022-01-11T08:39:57Z","publication":"IEEE/IFIP Network Operations and Management Symposium (NOMS)","keyword":["network management","service management","AI","Monte Carlo Tree Search","model-based","QoS"],"quality_controlled":"1","author":[{"last_name":"Werner","first_name":"Stefan","full_name":"Werner, Stefan"},{"orcid":"0000-0001-8210-4011","full_name":"Schneider, Stefan Balthasar","first_name":"Stefan Balthasar","id":"35343","last_name":"Schneider"},{"last_name":"Karl","id":"126","first_name":"Holger","full_name":"Karl, Holger"}],"publisher":"IEEE","file":[{"access_level":"open_access","date_created":"2022-01-11T08:39:57Z","file_name":"author_version.pdf","date_updated":"2022-01-11T08:39:57Z","content_type":"application/pdf","relation":"main_file","file_size":528653,"creator":"stschn","file_id":"29222"}],"conference":{"name":"IEEE/IFIP Network Operations and Management Symposium (NOMS)","start_date":"2022-04-25","location":"Budapest","end_date":"2022-04-29"},"_id":"29220","citation":{"apa":"Werner, S., Schneider, S. B., & Karl, H. (2022). Use What You Know: Network and Service Coordination Beyond Certainty. IEEE/IFIP Network Operations and Management Symposium (NOMS). IEEE/IFIP Network Operations and Management Symposium (NOMS), Budapest.","ama":"Werner S, Schneider SB, Karl H. Use What You Know: Network and Service Coordination Beyond Certainty. In: IEEE/IFIP Network Operations and Management Symposium (NOMS). IEEE; 2022.","chicago":"Werner, Stefan, Stefan Balthasar Schneider, and Holger Karl. “Use What You Know: Network and Service Coordination Beyond Certainty.” In IEEE/IFIP Network Operations and Management Symposium (NOMS). IEEE, 2022.","bibtex":"@inproceedings{Werner_Schneider_Karl_2022, title={Use What You Know: Network and Service Coordination Beyond Certainty}, booktitle={IEEE/IFIP Network Operations and Management Symposium (NOMS)}, publisher={IEEE}, author={Werner, Stefan and Schneider, Stefan Balthasar and Karl, Holger}, year={2022} }","mla":"Werner, Stefan, et al. “Use What You Know: Network and Service Coordination Beyond Certainty.” IEEE/IFIP Network Operations and Management Symposium (NOMS), IEEE, 2022.","short":"S. Werner, S.B. Schneider, H. Karl, in: IEEE/IFIP Network Operations and Management Symposium (NOMS), IEEE, 2022.","ieee":"S. Werner, S. B. Schneider, and H. Karl, “Use What You Know: Network and Service Coordination Beyond Certainty,” presented at the IEEE/IFIP Network Operations and Management Symposium (NOMS), Budapest, 2022."},"year":"2022","type":"conference"},{"abstract":[{"lang":"eng","text":"Datacenter applications have different resource requirements from network and developing flow scheduling heuristics for every workload is practically infeasible. In this paper, we show that deep reinforcement learning (RL) can be used to efficiently learn flow scheduling policies for different workloads without manual feature engineering. Specifically, we present LFS, which learns to optimize a high-level performance objective, e.g., maximize the number of flow admissions while meeting the deadlines. The LFS scheduler is trained through deep RL to learn a scheduling policy on continuous online flow arrivals. The evaluation results show that the trained LFS scheduler admits 1.05x more flows than the greedy flow scheduling heuristics under varying network load."}],"ddc":["000"],"user_id":"63288","author":[{"first_name":"Asif","full_name":"Hasnain, Asif","last_name":"Hasnain","id":"63288"},{"full_name":"Karl, Holger","first_name":"Holger","id":"126","last_name":"Karl"}],"publisher":"IEEE Computer Society","publication":"2021 IEEE 18th Annual Consumer Communications & Networking Conference (CCNC)","keyword":["Flow scheduling","Deadlines","Reinforcement learning"],"status":"public","date_created":"2020-10-19T14:27:17Z","_id":"20125","conference":{"start_date":"2021-01-09","name":"2021 IEEE 18th Annual Consumer Communications & Networking Conference (CCNC)","location":"Las Vegas, USA","end_date":"2021-01-12"},"main_file_link":[{"url":"https://ieeexplore.ieee.org/document/9369514"}],"type":"conference","year":"2021","citation":{"short":"A. Hasnain, H. Karl, in: 2021 IEEE 18th Annual Consumer Communications & Networking Conference (CCNC), IEEE Computer Society, n.d.","ieee":"A. Hasnain and H. Karl, “Learning Flow Scheduling,” in 2021 IEEE 18th Annual Consumer Communications & Networking Conference (CCNC), Las Vegas, USA.","apa":"Hasnain, A., & Karl, H. (n.d.). Learning Flow Scheduling. In 2021 IEEE 18th Annual Consumer Communications & Networking Conference (CCNC). Las Vegas, USA: IEEE Computer Society. https://doi.org/10.1109/CCNC49032.2021.9369514","ama":"Hasnain A, Karl H. Learning Flow Scheduling. In: 2021 IEEE 18th Annual Consumer Communications & Networking Conference (CCNC). IEEE Computer Society. doi:https://doi.org/10.1109/CCNC49032.2021.9369514","chicago":"Hasnain, Asif, and Holger Karl. “Learning Flow Scheduling.” In 2021 IEEE 18th Annual Consumer Communications & Networking Conference (CCNC). IEEE Computer Society, n.d. https://doi.org/10.1109/CCNC49032.2021.9369514.","mla":"Hasnain, Asif, and Holger Karl. “Learning Flow Scheduling.” 2021 IEEE 18th Annual Consumer Communications & Networking Conference (CCNC), IEEE Computer Society, doi:https://doi.org/10.1109/CCNC49032.2021.9369514.","bibtex":"@inproceedings{Hasnain_Karl, title={Learning Flow Scheduling}, DOI={https://doi.org/10.1109/CCNC49032.2021.9369514}, booktitle={2021 IEEE 18th Annual Consumer Communications & Networking Conference (CCNC)}, publisher={IEEE Computer Society}, author={Hasnain, Asif and Karl, Holger} }"},"title":"Learning Flow Scheduling","department":[{"_id":"75"}],"publication_status":"accepted","project":[{"_id":"4","name":"SFB 901 - Project Area C"},{"_id":"16","name":"SFB 901 - Subproject C4"},{"name":"SFB 901","_id":"1"}],"date_updated":"2022-01-06T06:54:20Z","doi":"https://doi.org/10.1109/CCNC49032.2021.9369514","language":[{"iso":"eng"}]},{"user_id":"15504","title":"Automating Network Resource Allocation for Coflows with Deadlines","date_created":"2021-11-16T13:05:12Z","project":[{"_id":"1","name":"SFB 901"},{"name":"SFB 901 - Project Area C","_id":"4"},{"name":"SFB 901 - Subproject C4","_id":"16"}],"status":"public","department":[{"_id":"75"}],"author":[{"full_name":"Hasnain, Asif","first_name":"Asif","last_name":"Hasnain"}],"doi":"10.17619/UNIPB/1-1241 ","date_updated":"2022-01-06T06:57:40Z","_id":"27503","supervisor":[{"full_name":"Karl, Holger","first_name":"Holger","last_name":"Karl"}],"language":[{"iso":"eng"}],"type":"dissertation","citation":{"mla":"Hasnain, Asif. Automating Network Resource Allocation for Coflows with Deadlines. 2021, doi:10.17619/UNIPB/1-1241 .","bibtex":"@book{Hasnain_2021, title={Automating Network Resource Allocation for Coflows with Deadlines}, DOI={10.17619/UNIPB/1-1241 }, author={Hasnain, Asif}, year={2021} }","ama":"Hasnain A. Automating Network Resource Allocation for Coflows with Deadlines.; 2021. doi:10.17619/UNIPB/1-1241 ","apa":"Hasnain, A. (2021). Automating Network Resource Allocation for Coflows with Deadlines. https://doi.org/10.17619/UNIPB/1-1241 ","chicago":"Hasnain, Asif. Automating Network Resource Allocation for Coflows with Deadlines, 2021. https://doi.org/10.17619/UNIPB/1-1241 .","ieee":"A. Hasnain, Automating Network Resource Allocation for Coflows with Deadlines. 2021.","short":"A. Hasnain, Automating Network Resource Allocation for Coflows with Deadlines, 2021."},"year":"2021"},{"ddc":["000"],"user_id":"63288","abstract":[{"lang":"eng","text":"Data-parallel applications are developed using different data programming models, e.g., MapReduce, partition/aggregate. These models represent diverse resource requirements of application in a datacenter network, which can be represented by the coflow abstraction. The conventional method of creating hand-crafted coflow heuristics for admission or scheduling for different workloads is practically infeasible. In this paper, we propose a deep reinforcement learning (DRL)-based coflow admission scheme -- LCS -- that can learn an admission policy for a higher-level performance objective, i.e., maximize successful coflow admissions, without manual feature engineering. LCS is trained on a production trace, which has online coflow arrivals. The evaluation results show that LCS is able to learn a reasonable admission policy that admits more coflows than state-of-the-art Varys heuristic while meeting their deadlines."}],"status":"public","date_created":"2021-01-16T18:24:19Z","author":[{"first_name":"Asif","full_name":"Hasnain, Asif","last_name":"Hasnain","id":"63288"},{"last_name":"Karl","id":"126","first_name":"Holger","full_name":"Karl, Holger"}],"publisher":"IEEE Communications Society","publication":"IEEE INFOCOM 2021 - IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS)","keyword":["Coflow scheduling","Reinforcement learning","Deadlines"],"_id":"21005","conference":{"end_date":"2021-05-13","name":"IEEE INFOCOM 2021 - IEEE Conference on Computer Communications","start_date":"2021-05-10","location":"Vancouver BC Canada"},"type":"conference","year":"2021","citation":{"ieee":"A. Hasnain and H. Karl, “Learning Coflow Admissions,” in IEEE INFOCOM 2021 - IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS), Vancouver BC Canada.","short":"A. Hasnain, H. Karl, in: IEEE INFOCOM 2021 - IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS), IEEE Communications Society, n.d.","mla":"Hasnain, Asif, and Holger Karl. “Learning Coflow Admissions.” IEEE INFOCOM 2021 - IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS), IEEE Communications Society, doi:10.1109/INFOCOMWKSHPS51825.2021.9484599.","bibtex":"@inproceedings{Hasnain_Karl, title={Learning Coflow Admissions}, DOI={10.1109/INFOCOMWKSHPS51825.2021.9484599}, booktitle={IEEE INFOCOM 2021 - IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS)}, publisher={IEEE Communications Society}, author={Hasnain, Asif and Karl, Holger} }","chicago":"Hasnain, Asif, and Holger Karl. “Learning Coflow Admissions.” In IEEE INFOCOM 2021 - IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS). IEEE Communications Society, n.d. https://doi.org/10.1109/INFOCOMWKSHPS51825.2021.9484599.","apa":"Hasnain, A., & Karl, H. (n.d.). Learning Coflow Admissions. In IEEE INFOCOM 2021 - IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS). Vancouver BC Canada: IEEE Communications Society. https://doi.org/10.1109/INFOCOMWKSHPS51825.2021.9484599","ama":"Hasnain A, Karl H. Learning Coflow Admissions. In: IEEE INFOCOM 2021 - IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS). IEEE Communications Society. doi:10.1109/INFOCOMWKSHPS51825.2021.9484599"},"main_file_link":[{"url":"https://ieeexplore.ieee.org/document/9484599"}],"title":"Learning Coflow Admissions","related_material":{"link":[{"url":"https://ieeexplore.ieee.org/document/9484599","relation":"confirmation"}]},"publication_status":"accepted","project":[{"name":"SFB 901 - Subproject C4","_id":"16"},{"_id":"4","name":"SFB 901 - Project Area C"},{"name":"SFB 901","_id":"1"}],"department":[{"_id":"75"}],"doi":"10.1109/INFOCOMWKSHPS51825.2021.9484599","date_updated":"2022-01-06T06:54:42Z","language":[{"iso":"eng"}]},{"abstract":[{"text":"Services often consist of multiple chained components such as microservices in a service mesh, or machine learning functions in a pipeline. Providing these services requires online coordination including scaling the service, placing instance of all components in the network, scheduling traffic to these instances, and routing traffic through the network. Optimized service coordination is still a hard problem due to many influencing factors such as rapidly arriving user demands and limited node and link capacity. Existing approaches to solve the problem are often built on rigid models and assumptions, tailored to specific scenarios. If the scenario changes and the assumptions no longer hold, they easily break and require manual adjustments by experts. Novel self-learning approaches using deep reinforcement learning (DRL) are promising but still have limitations as they only address simplified versions of the problem and are typically centralized and thus do not scale to practical large-scale networks.\r\n\r\nTo address these issues, we propose a distributed self-learning service coordination approach using DRL. After centralized training, we deploy a distributed DRL agent at each node in the network, making fast coordination decisions locally in parallel with the other nodes. Each agent only observes its direct neighbors and does not need global knowledge. Hence, our approach scales independently from the size of the network. In our extensive evaluation using real-world network topologies and traffic traces, we show that our proposed approach outperforms a state-of-the-art conventional heuristic as well as a centralized DRL approach (60% higher throughput on average) while requiring less time per online decision (1 ms).","lang":"eng"}],"ddc":["000"],"user_id":"35343","publisher":"IEEE","author":[{"first_name":"Stefan Balthasar","orcid":"0000-0001-8210-4011","full_name":"Schneider, Stefan Balthasar","last_name":"Schneider","id":"35343"},{"full_name":"Qarawlus, Haydar","first_name":"Haydar","last_name":"Qarawlus"},{"last_name":"Karl","id":"126","first_name":"Holger","full_name":"Karl, Holger"}],"file_date_updated":"2021-03-18T17:12:56Z","publication":"IEEE International Conference on Distributed Computing Systems (ICDCS)","keyword":["network management","service management","coordination","reinforcement learning","distributed"],"file":[{"file_size":606321,"title":"Distributed Online Service Coordination Using Deep Reinforcement Learning","file_id":"21544","creator":"stschn","content_type":"application/pdf","date_updated":"2021-03-18T17:12:56Z","relation":"main_file","date_created":"2021-03-18T17:12:56Z","file_name":"public_author_version.pdf","access_level":"open_access"}],"status":"public","has_accepted_license":"1","date_created":"2021-03-18T17:15:47Z","_id":"21543","conference":{"name":"IEEE International Conference on Distributed Computing Systems (ICDCS)","location":"Washington, DC, USA"},"citation":{"chicago":"Schneider, Stefan Balthasar, Haydar Qarawlus, and Holger Karl. “Distributed Online Service Coordination Using Deep Reinforcement Learning.” In IEEE International Conference on Distributed Computing Systems (ICDCS). IEEE, 2021.","ama":"Schneider SB, Qarawlus H, Karl H. Distributed Online Service Coordination Using Deep Reinforcement Learning. In: IEEE International Conference on Distributed Computing Systems (ICDCS). IEEE; 2021.","apa":"Schneider, S. B., Qarawlus, H., & Karl, H. (2021). Distributed Online Service Coordination Using Deep Reinforcement Learning. In IEEE International Conference on Distributed Computing Systems (ICDCS). Washington, DC, USA: IEEE.","bibtex":"@inproceedings{Schneider_Qarawlus_Karl_2021, title={Distributed Online Service Coordination Using Deep Reinforcement Learning}, booktitle={IEEE International Conference on Distributed Computing Systems (ICDCS)}, publisher={IEEE}, author={Schneider, Stefan Balthasar and Qarawlus, Haydar and Karl, Holger}, year={2021} }","mla":"Schneider, Stefan Balthasar, et al. “Distributed Online Service Coordination Using Deep Reinforcement Learning.” IEEE International Conference on Distributed Computing Systems (ICDCS), IEEE, 2021.","short":"S.B. Schneider, H. Qarawlus, H. Karl, in: IEEE International Conference on Distributed Computing Systems (ICDCS), IEEE, 2021.","ieee":"S. B. Schneider, H. Qarawlus, and H. Karl, “Distributed Online Service Coordination Using Deep Reinforcement Learning,” in IEEE International Conference on Distributed Computing Systems (ICDCS), Washington, DC, USA, 2021."},"year":"2021","type":"conference","title":"Distributed Online Service Coordination Using Deep Reinforcement Learning","related_material":{"link":[{"relation":"software","url":"https://github.com/ RealVNF/distributed-drl-coordination"}]},"department":[{"_id":"75"}],"project":[{"_id":"1","name":"SFB 901"},{"_id":"4","name":"SFB 901 - Project Area C"},{"name":"SFB 901 - Subproject C4","_id":"16"}],"date_updated":"2022-01-06T06:55:04Z","oa":"1","language":[{"iso":"eng"}]},{"type":"conference","citation":{"bibtex":"@inproceedings{Schneider_Jürgens_Karl_2021, title={Divide and Conquer: Hierarchical Network and Service Coordination}, booktitle={IFIP/IEEE International Symposium on Integrated Network Management (IM)}, publisher={IFIP/IEEE}, author={Schneider, Stefan Balthasar and Jürgens, Mirko and Karl, Holger}, year={2021} }","mla":"Schneider, Stefan Balthasar, et al. “Divide and Conquer: Hierarchical Network and Service Coordination.” IFIP/IEEE International Symposium on Integrated Network Management (IM), IFIP/IEEE, 2021.","chicago":"Schneider, Stefan Balthasar, Mirko Jürgens, and Holger Karl. “Divide and Conquer: Hierarchical Network and Service Coordination.” In IFIP/IEEE International Symposium on Integrated Network Management (IM). IFIP/IEEE, 2021.","apa":"Schneider, S. B., Jürgens, M., & Karl, H. (2021). Divide and Conquer: Hierarchical Network and Service Coordination. In IFIP/IEEE International Symposium on Integrated Network Management (IM). Bordeaux, France: IFIP/IEEE.","ama":"Schneider SB, Jürgens M, Karl H. Divide and Conquer: Hierarchical Network and Service Coordination. In: IFIP/IEEE International Symposium on Integrated Network Management (IM). IFIP/IEEE; 2021.","ieee":"S. B. Schneider, M. Jürgens, and H. Karl, “Divide and Conquer: Hierarchical Network and Service Coordination,” in IFIP/IEEE International Symposium on Integrated Network Management (IM), Bordeaux, France, 2021.","short":"S.B. Schneider, M. Jürgens, H. Karl, in: IFIP/IEEE International Symposium on Integrated Network Management (IM), IFIP/IEEE, 2021."},"year":"2021","_id":"20693","conference":{"location":"Bordeaux, France","name":"IFIP/IEEE International Symposium on Integrated Network Management (IM)"},"author":[{"orcid":"0000-0001-8210-4011","full_name":"Schneider, Stefan Balthasar","first_name":"Stefan Balthasar","id":"35343","last_name":"Schneider"},{"first_name":"Mirko","full_name":"Jürgens, Mirko","last_name":"Jürgens"},{"id":"126","last_name":"Karl","full_name":"Karl, Holger","first_name":"Holger"}],"publisher":"IFIP/IEEE","quality_controlled":"1","file_date_updated":"2020-12-11T08:37:37Z","publication":"IFIP/IEEE International Symposium on Integrated Network Management (IM)","keyword":["network management","service management","coordination","hierarchical","scalability","nfv"],"file":[{"date_created":"2020-12-11T08:37:37Z","file_name":"preprint_with_header.pdf","access_level":"open_access","file_size":7979772,"file_id":"20694","creator":"stschn","title":"Divide and Conquer: Hierarchical Network and Service Coordination","content_type":"application/pdf","date_updated":"2020-12-11T08:37:37Z","relation":"main_file"}],"status":"public","has_accepted_license":"1","date_created":"2020-12-11T08:39:47Z","abstract":[{"lang":"eng","text":"In practical, large-scale networks, services are requested\r\nby users across the globe, e.g., for video streaming.\r\nServices consist of multiple interconnected components such as\r\nmicroservices in a service mesh. Coordinating these services\r\nrequires scaling them according to continuously changing user\r\ndemand, deploying instances at the edge close to their users,\r\nand routing traffic efficiently between users and connected instances.\r\nNetwork and service coordination is commonly addressed\r\nthrough centralized approaches, where a single coordinator\r\nknows everything and coordinates the entire network globally.\r\nWhile such centralized approaches can reach global optima, they\r\ndo not scale to large, realistic networks. In contrast, distributed\r\napproaches scale well, but sacrifice solution quality due to their\r\nlimited scope of knowledge and coordination decisions.\r\n\r\nTo this end, we propose a hierarchical coordination approach\r\nthat combines the good solution quality of centralized approaches\r\nwith the scalability of distributed approaches. In doing so, we divide\r\nthe network into multiple hierarchical domains and optimize\r\ncoordination in a top-down manner. We compare our hierarchical\r\nwith a centralized approach in an extensive evaluation on a real-world\r\nnetwork topology. Our results indicate that hierarchical\r\ncoordination can find close-to-optimal solutions in a fraction of\r\nthe runtime of centralized approaches."}],"ddc":["006"],"user_id":"35343","language":[{"iso":"eng"}],"date_updated":"2022-01-06T06:54:32Z","oa":"1","department":[{"_id":"75"}],"project":[{"name":"SFB 901","_id":"1"},{"name":"SFB 901 - Project Area C","_id":"4"},{"_id":"16","name":"SFB 901 - Subproject C4"}],"title":"Divide and Conquer: Hierarchical Network and Service Coordination"},{"language":[{"iso":"eng"}],"date_updated":"2022-01-06T06:55:15Z","doi":"10.1109/TNSM.2021.3076503","oa":"1","department":[{"_id":"75"}],"project":[{"_id":"1","name":"SFB 901"},{"_id":"4","name":"SFB 901 - Project Area C"},{"_id":"16","name":"SFB 901 - Subproject C4"}],"title":"Self-Learning Multi-Objective Service Coordination Using Deep Reinforcement Learning","type":"journal_article","citation":{"bibtex":"@article{Schneider_Khalili_Manzoor_Qarawlus_Schellenberg_Karl_Hecker_2021, title={Self-Learning Multi-Objective Service Coordination Using Deep Reinforcement Learning}, DOI={10.1109/TNSM.2021.3076503}, journal={Transactions on Network and Service Management}, publisher={IEEE}, author={Schneider, Stefan Balthasar and Khalili, Ramin and Manzoor, Adnan and Qarawlus, Haydar and Schellenberg, Rafael and Karl, Holger and Hecker, Artur}, year={2021} }","mla":"Schneider, Stefan Balthasar, et al. “Self-Learning Multi-Objective Service Coordination Using Deep Reinforcement Learning.” Transactions on Network and Service Management, IEEE, 2021, doi:10.1109/TNSM.2021.3076503.","chicago":"Schneider, Stefan Balthasar, Ramin Khalili, Adnan Manzoor, Haydar Qarawlus, Rafael Schellenberg, Holger Karl, and Artur Hecker. “Self-Learning Multi-Objective Service Coordination Using Deep Reinforcement Learning.” Transactions on Network and Service Management, 2021. https://doi.org/10.1109/TNSM.2021.3076503.","apa":"Schneider, S. B., Khalili, R., Manzoor, A., Qarawlus, H., Schellenberg, R., Karl, H., & Hecker, A. (2021). Self-Learning Multi-Objective Service Coordination Using Deep Reinforcement Learning. Transactions on Network and Service Management. https://doi.org/10.1109/TNSM.2021.3076503","ama":"Schneider SB, Khalili R, Manzoor A, et al. Self-Learning Multi-Objective Service Coordination Using Deep Reinforcement Learning. Transactions on Network and Service Management. 2021. doi:10.1109/TNSM.2021.3076503","ieee":"S. B. Schneider et al., “Self-Learning Multi-Objective Service Coordination Using Deep Reinforcement Learning,” Transactions on Network and Service Management, 2021.","short":"S.B. Schneider, R. Khalili, A. Manzoor, H. Qarawlus, R. Schellenberg, H. Karl, A. Hecker, Transactions on Network and Service Management (2021)."},"year":"2021","_id":"21808","publisher":"IEEE","author":[{"first_name":"Stefan Balthasar","orcid":"0000-0001-8210-4011","full_name":"Schneider, Stefan Balthasar","last_name":"Schneider","id":"35343"},{"first_name":"Ramin","full_name":"Khalili, Ramin","last_name":"Khalili"},{"last_name":"Manzoor","first_name":"Adnan","full_name":"Manzoor, Adnan"},{"first_name":"Haydar","full_name":"Qarawlus, Haydar","last_name":"Qarawlus"},{"full_name":"Schellenberg, Rafael","first_name":"Rafael","last_name":"Schellenberg"},{"id":"126","last_name":"Karl","full_name":"Karl, Holger","first_name":"Holger"},{"first_name":"Artur","full_name":"Hecker, Artur","last_name":"Hecker"}],"publication":"Transactions on Network and Service Management","file_date_updated":"2021-04-27T08:01:26Z","keyword":["network management","service management","coordination","reinforcement learning","self-learning","self-adaptation","multi-objective"],"file":[{"access_level":"open_access","file_name":"ris-accepted-version.pdf","date_created":"2021-04-27T08:01:26Z","content_type":"application/pdf","date_updated":"2021-04-27T08:01:26Z","description":"Author version of the accepted paper","relation":"main_file","file_size":4172270,"file_id":"21809","creator":"stschn"}],"status":"public","has_accepted_license":"1","date_created":"2021-04-27T08:04:16Z","article_type":"original","abstract":[{"text":"Modern services consist of interconnected components,e.g., microservices in a service mesh or machine learning functions in a pipeline. These services can scale and run across multiple network nodes on demand. To process incoming traffic, service components have to be instantiated and traffic assigned to these instances, taking capacities, changing demands, and Quality of Service (QoS) requirements into account. This challenge is usually solved with custom approaches designed by experts. While this typically works well for the considered scenario, the models often rely on unrealistic assumptions or on knowledge that is not available in practice (e.g., a priori knowledge).\r\n\r\nWe propose DeepCoord, a novel deep reinforcement learning approach that learns how to best coordinate services and is geared towards realistic assumptions. It interacts with the network and relies on available, possibly delayed monitoring information. Rather than defining a complex model or an algorithm on how to achieve an objective, our model-free approach adapts to various objectives and traffic patterns. An agent is trained offline without expert knowledge and then applied online with minimal overhead. Compared to a state-of-the-art heuristic, DeepCoord significantly improves flow throughput (up to 76%) and overall network utility (more than 2x) on realworld network topologies and traffic traces. It also supports optimizing multiple, possibly competing objectives, learns to respect QoS requirements, generalizes to scenarios with unseen, stochastic traffic, and scales to large real-world networks. For reproducibility and reuse, our code is publicly available.","lang":"eng"}],"ddc":["000"],"user_id":"35343"},{"_id":"33854","date_updated":"2022-11-18T09:59:27Z","oa":"1","year":"2021","citation":{"bibtex":"@book{Schneider_Karl_Khalili_Hecker_2021, title={DeepCoMP: Coordinated Multipoint Using Multi-Agent Deep Reinforcement Learning}, author={Schneider, Stefan Balthasar and Karl, Holger and Khalili, Ramin and Hecker, Artur}, year={2021} }","mla":"Schneider, Stefan Balthasar, et al. DeepCoMP: Coordinated Multipoint Using Multi-Agent Deep Reinforcement Learning. 2021.","chicago":"Schneider, Stefan Balthasar, Holger Karl, Ramin Khalili, and Artur Hecker. DeepCoMP: Coordinated Multipoint Using Multi-Agent Deep Reinforcement Learning, 2021.","apa":"Schneider, S. B., Karl, H., Khalili, R., & Hecker, A. (2021). DeepCoMP: Coordinated Multipoint Using Multi-Agent Deep Reinforcement Learning.","ama":"Schneider SB, Karl H, Khalili R, Hecker A. DeepCoMP: Coordinated Multipoint Using Multi-Agent Deep Reinforcement Learning.; 2021.","ieee":"S. B. Schneider, H. Karl, R. Khalili, and A. Hecker, DeepCoMP: Coordinated Multipoint Using Multi-Agent Deep Reinforcement Learning. 2021.","short":"S.B. Schneider, H. Karl, R. Khalili, A. Hecker, DeepCoMP: Coordinated Multipoint Using Multi-Agent Deep Reinforcement Learning, 2021."},"type":"working_paper","language":[{"iso":"eng"}],"abstract":[{"lang":"eng","text":"Macrodiversity is a key technique to increase the capacity of mobile networks. It can be realized using coordinated multipoint (CoMP), simultaneously connecting users to multiple overlapping cells. Selecting which users to serve by how many and which cells is NP-hard but needs to happen continuously in real time as users move and channel state changes. Existing approaches often require strict assumptions about or perfect knowledge of the underlying radio system, its resource allocation scheme, or user movements, none of which is readily available in practice.\r\n\r\nInstead, we propose three novel self-learning and self-adapting approaches using model-free deep reinforcement learning (DRL): DeepCoMP, DD-CoMP, and D3-CoMP. DeepCoMP leverages central observations and control of all users to select cells almost optimally. DD-CoMP and D3-CoMP use multi-agent DRL, which allows distributed, robust, and highly scalable coordination. All three approaches learn from experience and self-adapt to varying scenarios, reaching 2x higher Quality of Experience than other approaches. They have very few built-in assumptions and do not need prior system knowledge, making them more robust to change and better applicable in practice than existing approaches."}],"ddc":["004"],"title":"DeepCoMP: Coordinated Multipoint Using Multi-Agent Deep Reinforcement Learning","user_id":"477","author":[{"orcid":"0000-0001-8210-4011","full_name":"Schneider, Stefan Balthasar","first_name":"Stefan Balthasar","id":"35343","last_name":"Schneider"},{"last_name":"Karl","id":"126","first_name":"Holger","full_name":"Karl, Holger"},{"last_name":"Khalili","full_name":"Khalili, Ramin","first_name":"Ramin"},{"last_name":"Hecker","full_name":"Hecker, Artur","first_name":"Artur"}],"department":[{"_id":"75"}],"file_date_updated":"2022-10-20T16:41:10Z","keyword":["mobility management","coordinated multipoint","CoMP","cell selection","resource management","reinforcement learning","multi agent","MARL","self-learning","self-adaptation","QoE"],"file":[{"access_level":"open_access","file_name":"preprint.pdf","date_created":"2022-10-20T16:41:10Z","relation":"main_file","content_type":"application/pdf","date_updated":"2022-10-20T16:41:10Z","file_id":"33855","creator":"stschn","file_size":2521656}],"status":"public","has_accepted_license":"1","date_created":"2022-10-20T16:44:19Z","project":[{"name":"SFB 901 - C: SFB 901 - Project Area C","_id":"4"},{"_id":"16","name":"SFB 901 - C4: SFB 901 - Subproject C4"},{"_id":"1","name":"SFB 901: SFB 901"}]},{"oa":"1","_id":"35889","date_updated":"2023-01-10T15:09:05Z","type":"working_paper","citation":{"chicago":"Schneider, Stefan Balthasar. Conventional and Machine Learning Approaches for Network and Service Coordination, 2021.","ama":"Schneider SB. Conventional and Machine Learning Approaches for Network and Service Coordination.; 2021.","apa":"Schneider, S. B. (2021). Conventional and Machine Learning Approaches for Network and Service Coordination.","mla":"Schneider, Stefan Balthasar. Conventional and Machine Learning Approaches for Network and Service Coordination. 2021.","bibtex":"@book{Schneider_2021, title={Conventional and Machine Learning Approaches for Network and Service Coordination}, author={Schneider, Stefan Balthasar}, year={2021} }","short":"S.B. Schneider, Conventional and Machine Learning Approaches for Network and Service Coordination, 2021.","ieee":"S. B. Schneider, Conventional and Machine Learning Approaches for Network and Service Coordination. 2021."},"year":"2021","language":[{"iso":"eng"}],"title":"Conventional and Machine Learning Approaches for Network and Service Coordination","ddc":["004"],"user_id":"35343","abstract":[{"lang":"eng","text":"Network and service coordination is important to provide modern services consisting of multiple interconnected components, e.g., in 5G, network function virtualization (NFV), or cloud and edge computing. In this paper, I outline my dissertation research, which proposes six approaches to automate such network and service coordination. All approaches dynamically react to the current demand and optimize coordination for high service quality and low costs. The approaches range from centralized to distributed methods and from conventional heuristic algorithms and mixed-integer linear programs to machine learning approaches using supervised and reinforcement learning. I briefly discuss their main ideas and advantages over other state-of-the-art approaches and compare strengths and weaknesses."}],"has_accepted_license":"1","status":"public","date_created":"2023-01-10T15:08:50Z","project":[{"_id":"1","name":"SFB 901: SFB 901"},{"_id":"4","name":"SFB 901 - C: SFB 901 - Project Area C"},{"name":"SFB 901 - C4: SFB 901 - Subproject C4","_id":"16"}],"author":[{"first_name":"Stefan Balthasar","full_name":"Schneider, Stefan Balthasar","orcid":"0000-0001-8210-4011","last_name":"Schneider","id":"35343"}],"file_date_updated":"2023-01-10T15:07:03Z","department":[{"_id":"75"}],"keyword":["nfv","coordination","machine learning","reinforcement learning","phd","digest"],"file":[{"access_level":"open_access","date_created":"2023-01-10T15:07:03Z","file_name":"main.pdf","relation":"main_file","date_updated":"2023-01-10T15:07:03Z","content_type":"application/pdf","creator":"stschn","file_id":"35890","file_size":133340}]}]