Description du projet
Service and Software Architectures, Infrastructures and Engineering
Software is everywhere. And it needs maintenance: upgrades, fixes, new versions... But how can you be sure that a software installation will not break something which was working previously?
Did you ever install an upgrade on your PC, just to find out afterwards that something very important was not working anymore?
The problem is that there is no general way to know if a complex software package, with all its dependencies on obscure libraries, configuration files, hardware components or security setup, will work properly on a specific computer. I mean, there is no way to know it BEFORE actually installing it, and before finding out the hard way that it has destroyed something else.
MANCOOSI wants to find reliable solutions to this problem.
First of all, the project is studying a way to model accurately the installation process, in order to be able to know, without installing a piece of software, whether its installation will cause problems on a given configuration; this is not easy, and in complex cases can be a very difficult problem to solve. Understanding the theory behind the issue will allow the development of a new generation of tools, which will make software upgrading a simpler process for any user, not only for experienced computer wizards.
To test the theory, MANCOOSI is working on several versions of Linux, which are very complex pieces of software, and developing sophisticated optimization algorithms to find efficient upgrade paths – if you ever had to wait half an hour in front of a screen waiting for the "installation completed" message you know what I mean.
You can have a wonderful technology, but still something can go wrong. There will always be the possibility that the final result of the installation is not what the user really wanted; that's why MANCOOSI is also developing a "time machine", which will allow to bring your software system back to a previous state without further problems. It will be something similar to what is already available in some operating systems, but much more powerful: a tool to travel back to any moment in the past, restoring your system to a stable state with all the applications working exactly as they were at that time.
The MANCOOSI project is focused on Open Source software, but it is expected to deliver results which will be in principle applicable to any type of software. And since now many devices have some kind of software inside (telephones, cars, TVs, media players, washing machines, medical equipment, GPS …) we believe that MANCOOSI will have a big impact on the daily life of people, making software more robust, more reliable and user-friendly.
MANCOOSI project started on 31/01/2008, and is expected to deliver all its results by 31/01/2011.
Mancoosi aims to develop the scientific knowledge and build the tools necessary to manage the complexity of the open source infrastructure, which is one of the essential building blocks of tomorrow's software architectures: the success of LAMP (Linux, Apache, Mysql, Php) inside and outside the data centers is clear evidence of this. Yet, this infrastructure undergoes a fast-paced and distributed evolution that is too often maintained in ad-hoc ways using tools and processes that have clearly attained their limits today: we need new and innovative technology.
We explicitly target the difficult problems that arise when one wants to efficiently and safely upgrade a set of software components in complex software infrastructures, like those found in open source software distributions, among the most complex software systems known, made of tens of thousands of components that evolve over time without centralized design.
Mancoosi will provide: a model of the infrastructure, and the transformations it undergoes when adding or removing components; advanced algorithm to choose efficient evolution paths when updating a platform; a forum to attract leading experts by organizing an international competition; tools that incorporate these findings and advance the state of the art in the field.
Mancoosi is precisely focused on Objective ICT-2007.1.2 (c) of the present call, by enabling mastery of complexity, dependability and behavioral stability in the complex system of software infrastructures evolving over time without central design.
We bring together innovative industries from the Open Source world with deep first-hand experience in the practical issues of the problem, members of the Open Source community that are able to provide high-quality access to the community, and leading researchers that have the knowledge necessary to elaborate the sophisticated models and algorithms needed to solve the underlying problems.
Champ scientifique
Appel à propositions
FP7-ICT-2007-1
Voir d’autres projets de cet appel
Régime de financement
CP - Collaborative project (generic)Coordinateur
F-75205 / 13 Paris
France