1. How would you present the project?
Cloud computing is still one of the fastest growing market in computing industry. As such, newcomers in this market have to deal with a huge variety of providers and technologies, not only from a hosting point of view but also for developing « cloud ready » applications. Along with these technologies comes plethora of standards that I call « catalogs » : they define how to call a virtual machine, a Java platform and so on. Aside from the difficulty to choose between one standard or another, these standards are not able to embrace the growing variety of services that exists in the cloud computing domain. As more and more domains are converging toward a service delivery model – like internet of things, open data, etc – developers need standard tools and languages to describe interactions between these services, given that any standard will always be focused on an existing domain.
OCCIware objective is to identify the common denominator between all these services and provide an engineering toolchain for developping these services. We are concretly providing a meta-model and a language associated with model-driven engineering tools (Eclipse extension for instance), compilers, runtimes, libraries and so on. We base our work on an existing technology OCCI, that we will extend as necessary to handle many use cases beyond the sole cloud computing applications.
For us, the success of the project will be evaluated against its level of adoption by cloud computing. That's why we've tried fairies to lean over OCCIware's craddle and we can allready say the project is strongly supported by a standardization organization (Open Grid Forum) and a committee of 11 top scientists and experts in the domain of cloud computing.
2. What is your role in the project?
Since the beginning, I was contacted by Inria and TSP researchers to take part in the core team defining project objectives. Due to my past experience with OCCI technology and my involvement in the OCCI working group, I am contributing to various tasks of the project : extensions of the meta-model, architecture of the runtime, etc. I am also deeply involved in the dissemination actions through scientific publications and evangelization in industrial events. Finally, as chairman of the Strategic Orientation Committee, which gathers feedback from industrial and scientific community on the project, I am contributing in the global strategy of the project.
3. What key innovation do you bring or help to develop?
My main contribution so far is the erocci framework, a generic OCCI implementation written with the erlang/OTP platform. This project, hosted by OW2 and started in 2013, has already been awarded Best Innovation Project in 2014. But while the erlang/OTP platform brings unique features in term of scalability, resiliency, deployment and maintainability, most of enterprise applications use more mainstream languages (Java, Python, PHP, etc). In the frame of OCCIware project, we are now defining an architecture were erocci can serve as a core engine for a multi-language interconnection between existing applications or as an inspiration for other runtimes.
4. A word about yourself and your organization.
Since 2011, I have been working on OCCI technology, through the CompatibleOne project, participation in the OCCI working group, then developing the erocci project. Before that, I have mainly worked on distributed systems, in the HPC or cloud computing domain. I have always worked with open source projects, considering it a good model for sharing innovation, promoting excellence and personal career development as an engineer. Since 1 year now, I am particularly involved in the OW2 consortium as the representative of individual members and Technology Council chairman.
For this project, I have been hired by Inria in a team whose focus is on software and languages engineering. OCCIware brings interesting use cases regarding their work and we will lead the actions regarding the extensions of the meta-model, language definition and, in general, theoretical aspects of the project.