{"title":"Monte Carlo Tree Search for Cross-Stratum Optimization of Survivable Inter-Data Center Elastic Optical Network","authors":"Michal Aibin, K. Walkowiak","doi":"10.1109/RNDM.2018.8489841","DOIUrl":null,"url":null,"abstract":"In last few years, cloud computing and other services based on data centers have evolved from an emerging technology to a recognized approach that is gaining broad acceptance and deployment. Therefore, there is a significant need to provide efficient and reliable operation of inter-data center networks based on optical technologies. In this article, we focus on cross stratum optimization of an inter-data center elastic optical network with additional survivability requirements. We propose a novel optimization approach that employs machine learning Monte Carlo Tree Search (MCTS) algorithm for simulation of future traffic to improve the performance of the network regarding blocking probability and operational cost. We evaluate the performance of the proposed method and other reference algorithms under various network scenarios, using representative topologies and real data provided by Amazon Web Services. The main conclusion is that the approach based on the MCTS algorithm enables better coordination of resource allocation in both strata, which results in lower blocking of requests and lower cost taking into account the survivability requirements, in comparison to other algorithms.","PeriodicalId":340686,"journal":{"name":"2018 10th International Workshop on Resilient Networks Design and Modeling (RNDM)","volume":"51 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2018-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"4","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2018 10th International Workshop on Resilient Networks Design and Modeling (RNDM)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/RNDM.2018.8489841","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 4
Abstract
In last few years, cloud computing and other services based on data centers have evolved from an emerging technology to a recognized approach that is gaining broad acceptance and deployment. Therefore, there is a significant need to provide efficient and reliable operation of inter-data center networks based on optical technologies. In this article, we focus on cross stratum optimization of an inter-data center elastic optical network with additional survivability requirements. We propose a novel optimization approach that employs machine learning Monte Carlo Tree Search (MCTS) algorithm for simulation of future traffic to improve the performance of the network regarding blocking probability and operational cost. We evaluate the performance of the proposed method and other reference algorithms under various network scenarios, using representative topologies and real data provided by Amazon Web Services. The main conclusion is that the approach based on the MCTS algorithm enables better coordination of resource allocation in both strata, which results in lower blocking of requests and lower cost taking into account the survivability requirements, in comparison to other algorithms.
在过去几年中,基于数据中心的云计算和其他服务已经从一种新兴技术发展成为一种得到广泛接受和部署的公认方法。因此,为基于光技术的数据中心间网络提供高效、可靠的运行是迫切需要的。在本文中,我们关注具有额外生存性要求的数据中心间弹性光网络的跨层优化。我们提出了一种新的优化方法,采用机器学习蒙特卡罗树搜索(MCTS)算法来模拟未来的流量,以提高网络在阻塞概率和运行成本方面的性能。我们使用具有代表性的拓扑和Amazon Web Services提供的真实数据,评估了所提出的方法和其他参考算法在各种网络场景下的性能。主要结论是,与其他算法相比,基于MCTS算法的方法可以更好地协调两个层的资源分配,从而减少请求阻塞,降低成本,同时考虑到生存性要求。