Section: New Results

Application domain results

An emerging trend into Amazones team is to apply our northBound/southBound approach to the InternetOfThing wave. We try to apply our architectures to the IoT application domain.

Data Centric Applications Distribution

Peer to peer systems have been widely used to alleviate the burden of servers by transferring to peers in a network tasks that do not require a centralization of the information. A wide range of applications are now emerging over peer-to-peer, such as social networking, multiplayer games, mobile messaging, etc. Most of these applications are essentially data-centric, they rely on exchange of data between peers, and can be expressed by queries over the database.

We propose a tool that allows for such applications, programmed as a collection of queries over a database, to be ported seamlessly without changing the initial queries from a client/server system to peer-to-peer system. The distribution is done with overlays network defined by declarative data centric programs specified in the Netlog language, thus resulting in a fully data centric modeling of the peer-to-peer application. The communication between peers relies on implicit addresses which can be evaluated on the fly to ensure the persistence of data.

We demonstrate the technique on a multiplayer online game, written in SQL, with players who connect to a mobile ad hoc network through their portable devices. The overlay is defined by a combination of an ad hoc routing protocol, DSDV, together with a DHT. The application runs on the QuestMonitor system, which allows to monitor the communication between peers, the evolution of the local data stores, as well as the execution of the declarative code.

Service Deployment in Disrupted Networks

OLD / REMOVE ? Ambient environments classically use wireless connections that suffers from frequent disconnections. The hard research point is to ensure service continuity. This disconnection problem has been widely tackled for application data with proxy and prefetching approaches. For services, disconnections are more difficult to anticipate, since service calls are only solved at run-time.

We are currently working on service deployment and invocations in disrupted networks with a network coding approach. This research is a joint work with the Swing team and with Aline Viana (INRIA Saclay @ TU Berlin). The main idea is to study how social-oriented applications, that need inter-dependent services and updates to be distributed to all or part of the mobile users community, could benefit from a network coding approach. The project aims at assessing for the first time the performance, in terms of latency, energy efficiency and capacity, of standard network coding techniques in presence of realistic user mobility and service demands. Building on these results, we plan to propose original social-aware network coding techniques that take advantage of the heterogeneous nature of the opportunistic network to reduce delays and energy consumption, in presence of multiple concurrent service flows targeting either all users or specific groups of interests.

These performance issues tackle at the same time the overall network capacity optimizations, as well as the overall software stack optimizations of a device with local and autonomous network coding strategies.

An INRIA ARC project proposal, entitled SoCool, has been submitted joinly with INRIA AMAZONES, INRIA SWING, INRIA MAESTRO, INRIA Saclay, University of Nice, TU Berlin and Fordham University.