Abstract
Virtual network mapping (VNM) is a challenge in the field of network virtualization. As VNM variants have been formalized depending on substrate network structures, virtual network specifications, mapping optimization objectives, and other factors, a number of VNM heuristic methods have been introduced. On the other hand, reinforcement learning (RL) algorithms have been incorporated into deep learning frameworks and recognized as a promising solution for solving complex resource allocation problems. In this paper, we present an RL-based graph embedding and mapping framework, Gemma, for tackling various VNM problems in a unified end-to-end manner. In the framework, we employ an encoder-decoder deep learning architecture and propose several optimization schemes such as two-stage mapping and model-based selective embedding. Aiming to deal with large-scale VNM problems in both online and offline scheduling systems, the proposed schemes explore the trade-off between inference accuracy and mapping function runtimes, enhancing scalability and timeliness. Gemma shows robust performance under various problem conditions, outperforming other heuristic and learning-based methods.
Highlights
In network virtualization, virtual network mapping (VNM) or virtual network embedding is a fundamental function by which virtual networks with resource requests on computing nodes and communication links are mapped onto a substrate network
In [12], VNM problems were formalized as an Markov decision process (MDP) where continual decisions about selecting target substrate nodes for given virtual nodes are made, and those problems were addressed through reinforcement learning (RL) such as the Q-learning algorithm
We present a generic VNM framework, Gemma that can deal with various VNM problems by exploiting neural graph embedding
Summary
Virtual network mapping (VNM) or virtual network embedding is a fundamental function by which virtual networks with resource requests on computing nodes and communication links are mapped onto a substrate network. VNM enables a network operator or data center manager to provision several virtual networks on a single substrate network infrastructure, aiming to utilize underlying resources more efficiently and to maximize the benefits achieved from the infrastructure. In [12], VNM problems were formalized as an Markov decision process (MDP) where continual decisions about selecting target substrate nodes for given virtual nodes are made, and those problems were addressed through reinforcement learning (RL) such as the Q-learning algorithm These prior works did not fully explore the advantages and complexity of using neural graph embedding to deal with complex virtual network structures or establish fundamental design principles for large scale VNM problems. We employ a model-based selective embedding scheme that can reduce the computation loads of graph embedding tasks for a time-varying network This scheme exploits a prediction model about graph embedding changes, i.e., the difference between a pair of embedding data corresponding to before and after the mapping of a virtual node.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.