Evaluation of the abstraction of optical topology models in blockchain-based data center interconnection

【Author】 Alemany, Pol; Vilalta, Ricard; Munoz, Raul; Casellas, Ramon; Martinez, Ricardo

【Source】JOURNAL OF OPTICAL COMMUNICATIONS AND NETWORKING

【影响因子】4.508

【Abstract】Data center (DC) interconnection allows us to have optical transmissions between DCs directly connected to optical networks, avoiding the use of a packet-based infrastructure. Thanks to the use of next-generation pluggable coherent optics, it is possible to create connectivity services (CSs) across multiple optical transport domains. In this multi-domain CS scenario, cloud operators and transport operators have to work together in the most dynamic way possible. To do so, they need a common place (i.e., a market) where the transport operators may expose their available optical resources and the cloud operators request (e.g., rent) them to be used in order to create end-to-end (E2E) CSs between DCs. Having multiple transport operators exposing their resource information in a common place requires a set of common rules (i.e., how much of the topology to show) to create E2E CSs requested between cloud operators. This paper makes use of the blockchain technology to present a blockchain-based extension for the software-defined network (SDN) architecture to allow each optical transport operator domain to become a peer in a blockchain network. In there, each peer follows the same rules and shares the same exact level of topology information by using a specific abstraction model to map the optical domain resources. This paper uses a set of three different abstraction models to validate their behavior on a blockchain system when managing multiple domain resources and the deployment of CSs across these domains. To do so, an experimental comparison on how the different abstraction models affect the performance of the blockchain system is presented. (C) 2022 Optica Publishing Group

【Keywords】

【发表时间】2022 APR

【收录时间】2022-03-09

【文献类型】期刊

【主题类别】

区块链应用-网络安全-