Sensitive media analysis through deep learning architectures
Industry 4.0 system concepts and machine learning for intelligent environment design
Few-Shot Scene Recognition, a New Proxy Task for Child Sexual Abuse Material Detec...
| Grant number: | 20/05174-0 |
| Support Opportunities: | Research Grants - Innovative Research in Small Business - PIPE |
| Duration: | June 01, 2021 - May 31, 2023 |
| Field of knowledge: | Engineering - Electrical Engineering - Telecommunications |
| Convênio/Acordo: | MCTI/MC |
| Principal Investigator: | Epaminondas Aguiar de Sousa Junior |
| Grantee: | Epaminondas Aguiar de Sousa Junior |
| Host Company: | Vixphy - Soluções de Computação em Nuvem Ltda |
| CNAE: |
Consultoria em tecnologia da informação Suporte técnico, manutenção e outros serviços em tecnologia da informação |
| City: | Vitória |
| Pesquisadores principais: | Moisés Renato Nunes Ribeiro |
| Associated researchers: | Cristina Klippel Dominicini ; Giovanni Ventorim Comarela ; Leandro Colombi Resendo ; Magnos Martinello ; Renato Souza Silva ; Rodolfo da Silva Villaca ; Welton Sthel Duque |
| Associated scholarship(s): | 21/11518-7 - SAWI - Savvy Access through Worldwide Internet,
BP.TT 21/10909-2 - SAWI - Savvy Access through Worldwide Internet, BP.TT 21/10978-4 - SAWI - Savvy Access through Worldwide Internet, BP.TT 21/09198-4 - SAWI - Savvy Access through Worldwide Internet, BP.TT |
Abstract
In this research project, we addressed the problem of the programmability of overlapping network architectures. A network architecture can be defined as a distributed communication system having the following attributes: (i) Control plan that contains the control logic, impacting the design of the network algorithms; (ii) Data plan that makes the routing decisions (based on flow, a destination address, etc.); (iii) Management plan that deals with the state in which the network operates (for switching, routing or QoS); and (iv) Knowledge plan that deals with learning techniques, automation and intelligent orchestration; Network programmability allows network designers to add, remove, or modify the state of the network on demand. By dynamically configuring the states on network elements, designers can architect their networks so that they behave optimally according to some performance objective (e.g., low latency or high flow). This can be achieved in various ways such as by programming the intermediate nodes of the network, by programming the routing, or even by managing the flows directed by algorithms to ensure the services offered to the final systems. Network programmability is a challenging problem given the complexity of defining a generic programmable network model and a set of programming interfaces that include routing, switching, access control, and routing. Despite technological advances in SDN, NFV, and Cloud, to form the foundations of modern networks, their functions, and their computing, there are a number of challenges that need to be addressed, especially to efficiently manage WAN networks and ensure dependability in the connectivity service. In this project, we propose as a transformative step to advance the state of the art, the construction of a platform called SAWI (Savvy Access through Worldwide Internet) as an enabling "Virtual Access Providers" to the Internet. The SAWI approach differentiates itself from existing solutions by exploring network programmability together with virtual programmable elements in hybrid clouds, in order to allow architectural elements to be orchestrated from machine learning techniques (self-driving network). (AU)
| Articles published in Agência FAPESP Newsletter about the research grant: |
| More itemsLess items |
| TITULO |
| Articles published in other media outlets ( ): |
| More itemsLess items |
| VEICULO: TITULO (DATA) |
| VEICULO: TITULO (DATA) |