The Discovery Initiative

Discover what could happen beyond the Clouds !

An Open-Science Initiative aiming at implementing a fully decentralized IaaS manager

The Discovery initiative aims to overcome the main limitations of the traditional server-centric cloud solutions by revising the OpenStack software in order to make it inherently cooperative.

Open positions !

PhD grant@Orange on Distributed SDN mechanisms (Subject in French)
Postdoc grant@Inria on Sandboxing technologies within the OpenStack ecosystem (baremetals, containers and VMs)
Phd grant@Inria on Reconfiguration of large scale distributed systems and software in a Fog computing context


The initiative aims at delivering a new generation of Utility Computing (UC) that will directly leverage any network point of presences (PoP, i.e.,a small or medium-sized network center) available through the Internet.

Fully Distributed

Implementing a fully Distributed IaaS manager is critical for the emergence of such a global infrastructure. Instead of developing from scratch, we decided to revise OpenStack with peer to peer mechanisms.

Fault Tolerant

``At large scale, failures become the norm rather than the exception'' The system we envision will natively deal with network partitioning concerns, enabling distinct parts of our global infrastructure to run independently (similarly to the Internet).


While the initiative focuses on OpenStack, secondary projects are also implemented in order to make progresses toward on goal. Major ones are referenced below.


A set of scripts leveraging the kolla and rally to conduct performance evaluations of OpenStack in a reproducible manner.

Get the Technical Report

See the OpenStack summit presentation

OpenStack Nova

Non intrusive fork of the Nova vanilla code (the MySQL DB has been replaced by the noSQL system REDIS). This works has been presented in the OpenStack Austin summit (April 2016) and the IEEE IC2E 2017 Conference

Get the technical report

See a demo


A load injector leveraging simigrid to perform large-scale simulations on VM management systems.

Get the paper