Lien vers le site de la Summer School Cloud Computing
Cloud Computing is a delivery model for IT, platform and infrastructure services based on Internet protocols. It typically involves provisioning of dynamically scalable and often virtualized resources at the infrastructure, platform and software levels. It addresses different fundamentals like (formal) modelling, scalability, optimisation, interoperability, quality of service, etc. This summer school will provide a forum for PhD students, scientists and technology leaders, experts and novices to facilitate knowledge sharing and information exchange on Cloud computing. It will promote fruitful discussions on the latest research and technology developments, directions, problems, and requirements.
The technical program covers the following topics:
Cloud networking
Cloud infrastructure
Cloud platforms
Cloud security
Cloud federationThe school was organised into several sessions, some of them consist of expert presentations from academia which cover the fundamental and theoretical aspects of the topics and from industry which cover latest technology developments and requirements. While other sessions consist of the hands-on workshops and poster presentations, which give the participants an excellent opportunity to gain experiences of technology and tools.
Cloud Infrastructures: Overview
Speakers: Gregory CUNHA and Julien MARIN-FRISONROCHE
Abstract: Cloud computing refers to “providing dematerialized computing resources as service on self-demand to customers” in opposition to old-fashion computing considering computing resources as products you must own. The three main types of services are IaaS/PaaS and SaaS which provides respectively infrastructures/(execution)platforms/services. We will focus on an overview of cloud infrastructures (IaaS layer) which provides to customers dematerialized infrastructures (network/firewall capabilities, virtual machines, etc.). Finally we will illustrate the presentation through a small demonstration.
Back to the program
StratusLab: A IaaS Cloud Distribution Focusing on Simplicity
Speaker: Charles LOOMIS and Mohammed AIRAJ (CNRS/LAL)
Abstract: The StratusLab collaboration provides an Infrastructure as a Service (IaaS) cloud distribution, allowing data centres to install private, community, or public clouds for their users. The distribution focuses on simplicity: allowing quick, straightforward installation on commodity hardware as well as easy access by scientists and engineers. Having run a cloud service based on StratusLab for several years, we will present issues that have arisen and provide examples of scientific use of the infrastructure. This will be followed by a live demonstration of the main features of the StratusLab cloud distribution.
More information about this tutorial is here.
Back to the program
Elastic-R and the building blocks of the Cloud
Speaker: Karim Chine
Abstract: TCloud computing is the answer to the explosion of big data. While the cloud provides infinite scalability for storage, several questions remain partly or fully unanswered: “How will we analyse all this data?”. “How can we analyse it virtually?”. “How can we leverage the programmability and elasticity of the cloud infrastructure to enhance the flexibility and capabilities of the software tools we use?”. “Will we be able to produce and publish on top of models and data, analytical services and GUIs as easily as we blog?”. “How will we snapshot, make reproducible, undo and redo easily data transformations and analysis?”. “Will we be able to achieve software convergence and make our data analysis tools communicate and work for us in synergy?”. “How will we view and analyse data collaboratively and how will we share the produced artifacts?”.
Elastic-R aims to answer these questions. For the benefit of both Academia and Industry, the Elastic-R platform transforms Amazon EC2 into a ubiquitous collaborative environment for data analysis and computational research. It makes the acquisition, use and sharing of all the capabilities required for statistical computing, data mining and numerical simulation easier than ever: The cloud becomes a user friendly Google-Docs-like platform where all the artifacts of computing can be produced by any number of geographically distributed real-time collaborators and can be stored, published and reused. The tutorial will be an overview of the Elastic-R platform, the latest developments will be demonstrated and applications in Bioinformatics will be illustrated.
Back to the program
Cloud Platforms: Concepts, Definitions, Architectures, and Open Issues
Speaker: Samir Tata
Abstract: Cloud computing is a recent paradigm in information technology enabling an economic model for virtual resources provisioning. In this paradigm, there are three well discussed layers of services known as IaaS for Infrastructure as a Service, PaaS for Platform as a Service and SaaS for Software as a Service. Other XaaS terms are used nowadays to name different resources provided as services in the cloud. In this talk, we will consider Cloud platforms (PaaS layer). We will mainly focus on current architectures and their underlining models. We will in addition consider some open research issues related to cloud platform, e.g. deployment, elasticity and mobility.
Back to the program
R and the Cloud
Speaker: Karim Chine
Abstract: Cloud Computing is holding the promise of democratizing access to computing infrastructures and deeply impacting research and education. However, the question “How will we bring the Infrastructure-as-a-Service paradigm to the data scientist’s desk and to the statistics classroom?” has remained unanswered. The Elastic-R Software platform proposes new concepts and frameworks to address this question: R, Python, Matlab, Spreadsheets, etc. are made accessible as articulated, programmable and collaborative components within a virtual and immersive environment for scientific research and higher education.
Teachers can easily and autonomously prepare interactive R-based custom learning environments and share them like documents in Google Docs. They can use them in the classroom or remotely in a distant learning context. They can also associate them with on-line-courses. Students are granted seamless access to pre-prepared, controlled and traceable learning environments. They can share their R sessions to receive guidance from Teachers or to solve problems in collaboration. Costs may be hidden to the students by allowing them to access temporarily shared institution-owned resources or using tokens that a teacher can generate using institutional cloud accounts.
Scientists can easily use the cloud as a ubiquitous and scriptable collaborative environment for traceable and reproducible data analysis and computational research. The cloud becomes a user friendly Google-Docs-like platform where all the artifacts of computing can be produced by any number of geographically distributed real-time collaborators and can be stored, published and reused. Big data access and analysis are simplified and made accessible to wider range of research professional. Science Gateways (graphical user interfaces for data science; set of tools, applications, and data integrated via portals) are made R-scriptable and hence easy to create, publish and update on the fly from the R command line: their use becomes an intrinsic part of the process of programming with Data.
The presentation will give an overview of the synergies that exist between R and the state-of-the art cloud technologies. Elastic-R on Amazon’s public cloud will be demonstrated via real-world applications in education, in bioinformatics and in finance.
Back to the program
Beyond the Cloud, The Discovery Initiative
Speaker: Adrien LÚbre, Institut Mines-Télécom, Ecole des Mines de Nantes
Abstract: The DISCOVERY initiative aims at exploring a new way of operating Utility Computing (UC) resources. To overcome the lack of sustainable solutions that can accommodate the ever increasing demand for computing resources, a disruptive change in UC platforms is required: Locality has to be considered as a primary concern. Instead of the current trend consisting of building larger and larger data centers in few strategic locations, the discovery initiative proposes to leverage any facilities available through the Internet in order to deliver widely distributed UC platforms that can better match the geographical dispersal of users as well as the unending demand. Although it involves radical changes in the way resources are managed, leveraging computing resources close to the end-users is the only solution to deliver a new generation of UC platforms, highly efficient and sustainable. Critical to the emergence of such locality-based UC (LUC) platforms is the availability of appropriate operating mechanisms. The DISCOVERY initiative aims at designing and implementing a unified system driving the use of resources at an unprecedented scale by turning a complex and diverse infrastructure into a collection of abstracted computing facilities that is both easy to operate and reliable. Internet Service Providers (ISPs) as well as academic and private institutions in charge of operating a part of the internet network will be able to build an extreme-scale LUC platform with a limited additional cost by simply deploying the DISCOVERY system through their infrastructure. To reach such a goal, we propose a ground-breaking way of operating UC platforms that leverages autonomous, decentralized and self-organizing techniques.
After describing the objectives and the major challenges of the initiative, we will focus in this talk on the two first building blocks that enables to schedule hundred thousands of VMs in a dynamic and cooperative manner.
Adrien LĂšbre is an Associate Professor at the Ecole des Mines de Nantes since 2008. He received his Ph.D. from Grenoble Institute of Technologies in September 2006 and his M.S. degree from the University of Grenoble in 2003. Between 2006 and 2008, he held a postdoc position at INRIA in the framework of the XtreemOS EU IP project. His research interests are distributed and Internet computing. More precisely he investigates the interest of recent programming models (such as event, actor and component programming) in the design of new systems in charge of operating large-scale distributed infrastructures. he has participated to several international, European and national actions, including the on-going EU Marie Curie SCALUS ITN (http://www.scalus.eu) and the French ANR SONGS project (http://infra-songs.gforge.inria.fr/). Besides, he is strongly involved in the Grid’5000/Aladdin consortium where he has been chairing the virtualization working group for 5 years. he has taken part to several program committees of international workshops and conferences (such as the ACM Virtualization Technologies in Distributed Computing Workshop that he co-chairs since 2011).
Local Organisation
- Chair: Samir Tata, Télécom SudParis, Institut Mines-Télécom, Samovar UMR 5157 Members:
- CĂ©line Bourdais, CNRS, Samovar UMR 5157
- Walid Gaaloul, Télécom SudParis,Institut Mines-Télécom, Samovar UMR 5157
- Chantal Taconet, Télécom SudParis,Institut Mines-Télécom, Samovar UMR 5157
Scientific committee
- Chair: Bruno Defude, Télécom SudParis, Institut Mines-Télécom, Samovar UMR 5157
Members:
Walid Ben Ameur, Télécom SudParis, Institut Mines-Télécom, Samovar UMR 5157
- Omar Boucelma, University of Aix-Marseille, LSIS, UMR 7296
- Samia Bouzefrane, CNAM, France
- Lionel Brunie, LIRIS, University of Lyon 1
- Christine Collet, Grenoble Institute of Technology, UMR 5217
- Thierry Delot, University of Valenciennes, UMR 8201
- Khalil Drira, LAAS – CNRS
- Mohamed Hamdi, SupâCom, Tunisie
- Marc Lacoste, Orange Labs, France
- Maryline Laurent, Télécom SudParis, Institut Mines-Télécom, Samovar UMR 5157
- Olivier Perrin, University of Lorraine, UMR 7503
- Michel Riveill, University of Nice – Sophia Antipolis
- Lionel Seinturier, University of Lille & Institut Universitaire de France
- Djamal Zeghlache, Télécom SudParis, Institut Mines-Télécom, Samovar UMR 5157