Design and evaluation of learning algorithms for dynamic resource management in virtual networks

Abstract

Abstract-Network virtualisation is considerably gaining attention as a solution to ossification of the Internet. However, the success of network virtualisation will depend in part on how efficiently the virtual networks utilise substrate network resources. In this paper, we propose a machine learning-based approach to virtual network resource management. We propose to model the substrate network as a decentralised system and introduce a learning algorithm in each substrate node and substrate link, providing self-organization capabilities. We propose a multiagent learning algorithm that carries out the substrate network resource management in a coordinated and decentralised way. The task of these agents is to use evaluative feedback to learn an optimal policy so as to dynamically allocate network resources to virtual nodes and links. The agents ensure that while the virtual networks have the resources they need at any given time, only the required resources are reserved for this purpose. Simulations show that our dynamic approach significantly improves the virtual network acceptance ratio and the maximum number of accepted virtual network requests at any time while ensuring that virtual network quality of service requirements such as packet drop rate and virtual link delay are not affected.

Contributions to Efficient Resource Management in Virtual Networks

“…Morever, as shown in 5, the better performance in RL-DRA is not at the expense of VN QoS, as the packet delay variation of the dynamic approach converges to that of the static one after the agent has learnt. We have written two papers [15], [16] as a result of these proposals. [15] is still under review while [16] has been accepted.…”

“…We have written two papers [15], [16] as a result of these proposals. [15] is still under review while [16] has been accepted.…”

Self-managed resources in network virtualisation environments

“…• A set of distributed learning-based self-management algorithms that allocate resources to virtual nodes and links dynamically, leading to better substrate resource utilisation [8], [11], [13], [14], [15], [16], [17], [18].…”

“…For a successful VNE, each virtual node must be mapped and any given substrate node can map at most one virtual node from the same request. Similarly, all the virtual links have to be mapped to one or more substrate links connecting the nodes to which the virtual nodes at its ends have been mapped without violating resource demand/availability constraints [8].…”

“…VNE involves the mapping of all nodes and links in a given VN to nodes and links respectively of a SN following a set of constraints [8]. The constrained VNE problem is NP-Hard [9].…”