RESEARCH PROJECTS

This pages presents the most recent research projects in which Prof. Joberto Martins participates.

COGCOM project develops the use of machine learning for the cognitive allocation of network communication resources for Internet of Things (IoT) Edge/Fog/Cloud distributed computing. 
The main technical and breakthrough emphasis of this research is the deployment of machine learning based systems as large data processing tool, thereby facilitating various services related to communication, transportation and energy utilization in smart cities.

REPAF Project - Dynamic and Cognitive Resource Allocation Model and Framework for MPLS, Elastic Optical Network (EON), Internet of Things (IoT) and Network Function Virtualization (NFV)

RePAF (Resource Provisioning and Allocation Framework) Project develops a new dynamic resource (LSP, bandwidth and lambda/ fibre) allocation solution computed  nearly on-the-fly with autonomic and cognitive characteristics.        RePAF develops a generalized resource allocation model (RPAM – Resource Provisioning and Allocation Model) based on Bandwidth Allocation Models (BAMs) and a framework (FDARM – Framework for Dynamic and Autonomic Resource Management) that integrates RPAM functionalities towards resource allocation with autonomic and cognitive characteristics.        FDRAM framework architecture is based on OpenFlow/SDN paradigm and integrates RPAM with the autonomic and cognitive functionalities by adopting  knowledge acquisition strategies and learning techniques for resource allocation.
The FIBRE testbed is as research facility constructed in the scope of a former project funded by the 2010 Brazil-EU Coordinated Call in ICT. It works as a large-scale Virtual Laboratory for students and researchers to test new applications and network architecture models. Currently, the FIBRE infrastructure consists of a federation of 11 local testbeds, also called “experimentation nodes” or simply “islands”. Each “island” have a set of network devices to support experiments in both fixed and wireless technologies. They are connected by an overlay network on the RNP backbone, comprised of two network separate layers: a control plane and an experiment plane. This network is named “FIBREnet”.