Programme Programme

Program of the Contrail Cloud Summer School 2012

Monday, July 23

09:30

Technical Tour Almere (optional)

This walking tour will guide you among several technology places related to Cloud technology in Almere. We will stop at:

-
  Cloud computing centre (tbc)
Almere houses some of the largests cloud data centres in the Netherlands. Here you will learn why they take the effort to put all the servers on the second and third floor in many Dutch data centres.
 
  AlmereGrid demo centre
AlmereGrid is involved in the Cloud test facility, but the main activity is operating a volunteer desktop grid with some 8.500 computers, which is connected with similar desktop grids into a pan-European Federation of about 500.000 computers. A volunteer desktop grid is a kind of cloud from the masses.
 
  Almere Smart City
Being a new town, Almere has many sensor systems in the city that turn the town into a Smart City. Some of the sensor networks, like the one in the bridge connecting the city to main land Netherlands, are analysed in HPC clouds. Developments are under way to include more of the networks in Sensor-as-a-Service clouds.
 
12:30-14:00 Lunch break  
14:00

Technical session: Cloud computing in the new town of Almere

Goal of this technical session is to get acquainted with Cloud computing efforts in and around the town of Almere. This as an illustration of practical work done in the deployment of Cloud technology. The session acts as a framework of reference for the scientific sessions later in the week.

 
14:00-14:30

The AlmereGrid Cloud Test Facility
Cloud computing infrastructures can be very complex with many components from many different vendors that all have to work together. AlmereGrid is setting up a Cloud Test Facility that acts as an independent test bed.

Download presentation

Ad Emmen (AlmereGrid)
14:30-15:00

The Dutch HealtHub
Big data is the next buzz word. The amounts of data  that need to be stored and processed in many scientific and data areas is growing exponentially. Coupling of data sets and reusing data is becoming more and more important. Mean while data security and data protection are growing in importance. The Dutch Health Hub, which builds on the Almere Data Capital Data Warehouse, is setting up a "Community Cloud" that brings together all players in the Dutch Health sector, from academic hospitals to service companies that work on or with big data in Dutch Healthcare.

Link to presentation
 

Peter Walgemoed (Dutch Health Hub)
15:00-15:30

WeNMR: bringing Grid computing to a worldwide structural biology community.
WeNMR brings together research teams in the Structural Biology area into a Virtual Research Community (VRC), focusing primarily on biomolecular Nuclear Magnetic Resonance (NMR) and Small Angle X-ray Scattering (SAXS). The WeNMR platform integrates and streamlines the computational approaches necessary for NMR and SAXS data analysis and structural modelling making use of a grid-based e-Infrastructure fully integrated into EGI (www.egi.eu ). Web portals provide user-friendly access to the grid, shielding end users from the complexity of the grid. The WeNMR Virtual Research Community has grown to become the largest in the life sciences area, accounting for about 33% of CPU usage within the life sciences area. WeNMR aims at serving all relevant communities around the world, providing a global e-Infrastructure platform and science gateway for structural biology in general. Of all portals currently in use, HADDOCK is responsible for about 16% of the WeNMR CPU usage and and growing quickly. In order to keep on serving our users with a fast calculation infrastructure we aim to expand the resources at our disposal. This involves not only additional EU Grid sites but also expansion into the field of desktop computing grids.

Download presentation

Marc van Dijk (WENMR)
15:30-16:00 Coffee/tea break  
16:00-16:30

The SARA/Big Grid HPC Cloud
Traditional Cloud infrastructures, like EC2 or Azure, are not very well suited for High Performance Computing (HPC). That is why SARA/Big Grid have started some time ago with a special HPC Cloud based on Open Nebula cloud middleware. It is now recognized as one of the world leaders in HPC Clouds and used as an example to copy in many other places around the world.

Download presentation

Tom Visser (SARA HPC Cloud)
16:30-17:00

Almere Smart Society - Spontaneous order from below or mass control from above?
Towards a Almere Smart Society; Amere Smart Society is a green, safe and vital urban environment of vibrant, diverse communities with healthy and sustainable systems. The citizen and the city use the full potential of intelligent digital infrastructures and ICT. These intelligent infrastructures connect people, business and city, endorse systemic and transsectoral collaboration, drive innovation, and simplify the co-creation of newservices. Almere Smart Society is a City of People with healthy economic growth, strong social cohesion, sustainable organic urban development and efficient city operations.'

Download presentation

Raymond Versteegh (Program Manager Consortium Almere Smart Society)

 

Tuesday, July 24

8.45-9:00 Opening Thilo Kielmann, Guillaume Pierre (Vrije Universiteit Amsterdam)
9:00-10:00

Contrail use cases and requirements

This talk discusses four challenging cloud applications that are being developed in the Contrail project. These applications make use of the various features of Contrail, and serve as validation use cases.

  • Distributed provision of geo-referenced data: a virtual 3D representation of the earth accessible by end users through a Web Control to be installed into the web browser. Users will be able to access maps, locations of services and other related services
  • Multimedia processing service marketplace: a web portal for the delivery of multimedia services, enabling end users to access contents adapted to their preferences or their devices.
  • Scientific Data Analysis: high volume data benig streamed from a large scientific experiment and analysed in real time; specifically data from the Small Angle Neutron Scattering (SANS) experiment.
  • Electronic Drug Discovery: use the power of the cloud to develop new drugs. Specifically, an open source programme called Bioconductor will be used to illustrate the power of the Contrail software stack.

Download presentation

 

Ad Emmen (Genias/AlmereGrid)
10:00-10:15 Coffee Break  
10:15-11:45

The open Cloud -- from mainframes to OpenStack

Download presentation

Darren Birkett (Rackspace inc.)
11:45-12:45 Lunch  
12:45-13:45

ConPaaS: an integrated runtime environment for elastic Cloud applications

Cloud computing offers a very flexible and cost-effective environment for hosting demanding applications. However, making use of these advanced functionalities can be tedious and error-prone: application developers must therefore handle the complexity of deploying applications composed of many inter-related components, implementing automatic resource provisioning, orchestrating application reconfigurations such that users do not notice any downtime, developing fault-tolerance mechanisms, etc. This presentation will introduce ConPaaS, an open-source runtime environment for hosting applications in the cloud which aims at offering the full power of the cloud to application developers while shielding them from the associated complexity of the cloud. ConPaaS is designed to host both high-performance scientific applications and online Web applications. It automates the entire life-cycle of an application, including collaborative development, deployment, performance monitoring, and automatic scaling. Finally, it runs on a variety of public and private clouds, and is easily extensible. This allows developers to focus their attention on application-specific concerns rather than on cloud-specific details.

Download presentation

Guillaume Pierre (Vrije Universiteit Amsterdam)
13:45-15:30

ConPaaS hands-on

In this session we will get hands-on experience with the ConPaaS system.

Download hands-on document

Guillaume Pierre (Vrije Universiteit Amsterdam)
15:30-15:45 Coffee Break  
15:45-16:45

XtreemFS

Cloud computing poses new challenges to data storage. While cloud providers use shared distributed hardware, which is inherently unreliable and insecure, cloud users expect their data to be safely and securely stored, available at any time and accessible in the same way as their locally stored data.

As a solution, we will present XtreemFS in this talk. XtreemFS is a POSIX-compliant file system for the cloud. It transparently supports advanced features like replication and distributed snapshots. A detailed description of the internal architecture of XtreemFS and its security concept will be given. Additionally, a comprehensive overview of the replication and snapshot protocols will be presented. The talk will conclude with a description how the presented XtreemFS features are used in practice at the example of a cloud provider.

Download presentation

Michael Berlin (Zuse Institut Berlin)

 

Wednesday, July 25

9:00-10:00

Task Farming in Contrail

Traditionally, high-performance computing has been about achieving high computing throughput by assembling as many machines as a researcher could get his hands on. In cloud computing, machines cost money per hour. Consequently, researchers need to care about the costs of their computations, and about cost efficiency. Questions like "which cloud offering gets my job done in time, within budget?" become even more important. But cloud providers do not have answers readily available.

In Contrail, the ConPaaS platform comes with a task farming service that addresses this issue. The service can compute bags of independent tasks, within a user-defined deadline and budget, without prior knowledge of task completion times of a particular bag of tasks, on given types of cloud infrastructure offerings. This lecture explains the fundamentals of cost-aware scheduling for clouds and the workings of ConPaaS' task farming service.

Download presentation

Thilo Kielmann (Vrije universiteit Amsterdam)
10:00-10:15 Coffee Break  
10:15-11:45

The Ibis e-Science Software Framework


In recent years, the application of high-performance and distributed computing in scientific practice has become increasingly widespread. Among the most widely available platforms to scientists are clusters, grids, and cloud systems. Such infrastructures currently are undergoing revolutionary change due to the integration of many-core technologies, providing orders-of-magnitude speed improvements for selected compute kernels. With high-performance and distributed computing systems thus becoming more heterogeneous and hierarchical, programming complexity is vastly increased. Further complexities arise because urgent desire for scalability, and issues including data distribution, software heterogeneity, and ad-hoc hardware availability, commonly force scientists into simultaneous use of multiple platforms (e.g. clusters, grids, and clouds used concurrently). A true computing jungle.

In this presentation we explore the possibilities of enabling efficient and transparent use of Jungle Computing Systems in every-day scientific practice. We discuss the fundamental methodologies required for defining programming models that are tailored to the specific needs of scientific researchers, e.g. as integrated in the Ibis high-performance distributed programming system. We also make a case for the urgent need for easy and efficient Jungle Computing in scientific practice, by exploring a set of state-of-the-art application domains, including computational astrophysics, multimedia analysis, and climate modelling.

Download presentation

Henri Bal (Vrije Universiteit)
11:45-12:45 Lunch  
12:45-14:15

TaskFarming Practical

In this session we will get hands-on experience with the TaskFarming service in ConPaaS.

Thilo Kielmann (Vrije universiteit Amsterdam)
14:15-14:30 Coffee Break  
14:30-16:45

PhD symposium

Tuning (Virtual) Machines - Naod D. Jebessa (University of Amsterdam)

Policy-Based Scheduling of Cloud Services - Faris Nizamic (University of Groningen)

On-Demand Service Adjustment for Mobile Usage using Cloud Techniques - Marvin Ferber (University of Bayreuth)

SLA-Driven Capacity Planning for Cloud Computing - Yousri Kouki (INRIA, France)

Leveraging Business Workflow Management for better, robust and rigorous large-scale experiment - Thomasz Buchert (INRIA&LORIA, France)

 

 

Thursday, July 26

8:45-10:15

Network Virtualization for Cloud Computing

Virtualization is a fundamental technique underlying Infrastructure-as-a-Service (Iaas) cloud computing that applies to the core sub-systems of a computer: processing, storage, and networking. While virtual machine (VM) technologies decouple and isolate processing within the scope of a single physical computer, virtual networking provide the basis for decoupling  and isolating communication among virtualized resources, enabling cloud infrastructures that can seamlessly span geographically-dispersed data centers, possibly across multiple institutions. This presentation overviews major software and hardware-based approaches to virtualizing network infrastructures at different layers – including packet capture/injection, switching and routing – with emphasis on technologies applicable to the dynamic establishment and management of user-level virtual networks for cloud and inter-cloud computing.
 
Renato Figueiredo (University of Florida)
10:15-10:30 Coffee Break  
10:30-11:30

Cloud Federations

The goal of this session is to introduce Cloud Federations, using Contrail as a reference example. We will discuss the issues of IaaS platforms spreading over several heterogeneous providers, as well as the options for exploiting these platforms, which range from brokering and cloud bursting to provisioning over multiple clouds.

Contrail IaaS federation assign a fundamental role to Service Level Agreements and their negotiation. In Contrail, SLA management at the federation level is closely linked to application and provider monitoring. As SLAs can express security-related constraints and continuously enforced ones, the Federation support needs mechanisms that go beyond simple interoperability. We will discuss these mechanisms in relation to multi-Cloud provisioning and application splitting.

Download presentation

Massimo Coppola (CNR, Italy)
11:30-12:30

SLA management

Service Level Agreements in Contrail adhere to a general model for hierarchical agreement negotiation developed within the SLA@SOI project. In this talk we will present the SLA model and basic elements and steps of SLA negotiations: SLA templates and negotiated SLAs, syntax and semantics of the SLA terms.

We will discuss and experiment with the current implementation as well as the plans for advanced SLA-related functionalities in Contrail. As specific characteristics of Contrail, which applies the hierarchical metaphor to allow multi-level negotiation between users, federation and the providers, we will introduce the distinction between specific and generic SLAs, the criteria for selecting providers and those for splitting SLAs in accordance to multi-provider deployment.

Download presentation

Massimo Coppola (CNR, Italy)

12:30-13:30 Lunch  
13:30-15:30

Federation practical

Download presentation

Massimo Coppola (CNR, Italy)
15:30-15:45 Coffee Break  
15:45-16:45

Federated identity management

Federated identity management is one of the hottest topics in research infrastructures: as researchers increasingly need to share resources to manage large data volumes or participate in multiple projects, easing the burden by providing single sign on and persistent identity management.  Although pushed strongly by academic institutions across the world, it is no less relevant for, say, the pharmaceutical industry. At the same time, cloud and online services providers such as Google and Yahoo provide OpenID identities which enable collaborations, and can be reused.  Building federations is a curious mix of technology, trust, federation policies, ease of use, user culture and habits, law, support, resource management and accounting, etc.  As all these have to work over a distributed infrastructure, more often than not crossing borders, security plays a strong role.  This presentation will primarily look at the available and emerging technologies, focusing on the pragmatic aspects: things that work in practice... Technologies covered include Shibboleth, Moonshot, credential conversion, the role of X.509 certificates, bootstrapping the federated infrastructure security, scalability, and delegation.

Download presentation

Download research paper

Jens Jensen (STFC, UK)

 

Friday, July 27

9:00-10:00

Contrail infrastructure

Download presentation

Piyush Harsh (INRIA, France)
10:00-11:30

Contrail infrastructure practical

Virtual Execution Platform is a key component of provider services of
the Contrail software stack. It interfaces the cloud provider's
resources with the federation and enables the deployment and full
lifecycle management of the OVF application on the provider's IaaS
resources. In this hands on session, you will learn about the RESTful
interface of VEP and how you can exploit it to create an user account,
create and manage OVF applications, manage application elasticity, etc.
You will also perform web-browser configuration to enable end-user's
access of the VEP's web-interface.

 

Piyush Harsh (INRIA, France)

 

11:30-12:00 Coffee Break  
12:00-13:00

Closing: the future of the cloud

Download presentation

Roberto Cascella (INRIA, France)
13:00-14:00 Lunch  

 

  

Summerschool 2012 Summerschool 2012

Contrail Summer School 2012

Cloud federation unscrambled

 

The  Cloud Computing Summer School was part of the Summer School Almere 2012 that took place in June en July 2012.

The Cloud computing  courses did place in week 30: 23-07-2012 - 27-07-2012.

 

Missed the 2012 Cloud computing Summerschool? You are welcome to our 2013 Cloud computing Summerschool.