Internship proposal (industrial research): Software-defined LAN modeling and classification Profile: Master 2 or engineering school
Supervisor: Ludovic Noirie (Alcatel-Lucent Bell Labs France)
E-Mail: [email protected], URL: https://www.bell-labs.com/researchers/584/
Location: Alcatel-Lucent Bell Labs France (Nozay, south of Paris).
Dates: Starting February/March 2015, duration 5~6 months
Alcatel Lucent is at the forefront of global communications, providing products and innovations in IP and cloud networking, as well as ultra-broadband fixed and wireless access. Researchers at Bell Labs collaborate closely with the company’s customers and product development teams to create the technologies that are transforming the way people connect with each other and with the information around them. The proposed internship takes place in the French location of the Bell Labs, within the Advanced Internet Research (AIR) department.
Because of the development of connected objects in number and variety, home-network management is an increasing burden for end-users. Within AIR, we are looking for an integrated software solution to the burden of home network management from end-users while improving and extending their digital experience. We call this solution the Majord’Home as it acts like a majordomo to control and manage the connected objects of the home and their relation with the external world [1].
For that purpose, we defined the Software-Defined LAN (SD-LAN) concept that can be generalized to other environments such as enterprise networks, smart buildings or smart cities. Following the Software- Defined Networking (SDN) principles, our solution relies on a virtualization layer that abstracts the Connected Objects (COs) belonging to the different users into Virtual Objects (VOs), in order to provide a programmable control of them through well-defined APIs. Our solution supports SD-LANs through the creation of service-oriented Communities of Connected Objects (CoCOs). The CoCO represents the end- points of the SD-LAN and embeds the attributes of COs in order to render the expected service. The COs of an SD-LAN are seen as belonging to the same LAN even if they belong to different environments.
The proposed internship is about preliminary studies on an autonomic solution to help the operators of the various smart environments in managing the SD-LANs that are established in their networks. The main objectives are the definition and the analysis of a solution to identify and classify the working SD-LANs within the network according to various criteria. The intern will:
1. Study and analyze the state of the art that is required for this study;
2. Define some requirements for autonomic management of SD-LANs in multiple environments;
3. Rework the CO, VO and CoCO modeling framework accordingly;
4. Define the abstract structure to encode the global view on all the working SD-LANs at the whole network scale (graph theory…);
5. Propose and analyze some first mechanisms to identify and classify SD-LANs (clustering, pattern matching…);
6. According to the internship progress, implement and validate these mechanisms (simulations).
We plan to have a follow-up of this work with a PhD thesis proposal on this thematic, consolidating the preliminary framework defined during the internship and exploiting it for self-discovery of new types of SD- LANs for service catalog enhancement, self-completion of SD-LAN requests to make the services working to help the users, self-diagnosis and self-healing of SD-LAN failures, etc.
References:
[1] Mathieu Boussard, Dinh Thai Bui, Richard Douville, Nicolas Le Sauze, Ludovic Noirie, Pierre Peloso, Rémi Varloot, Martin Vigoureux, The Majord'Home: a SDN Approach to Let ISPs Manage and Extend Their Customers' Home Networks, 1st International Workshop on Management of SDN and NFV Systems (ManSDN/NFV 2014, collocated with CNSM 2014), Nov 2014, Rio de Janeiro, Brazil. Paper available at: http://www.cnsm-conf.org/2014/proceedings/pdf/73.pdf.