PCS951 Cloud and Distributed Resources for High Volume Data Processing

Contents and structure

Access to high data volumes is a common feature of research projects in several fields of engineering and science. These amounts of data require novel approaches to extract key contents in a timely and efficient manner. Grid computing is an approach to handle heterogeneous processing in several distributed computing facilities. Cloud computing provides seamless access to remote facilities. The course concentrates on challenges related to safe and efficient utilisation of computing resources managed by heterogeneous operators, including protection concerns between the project owner and facility management. Software technology used for such systems is applied and configured.

This course covers technology and principles of grid and cloud computing, and gives practical introduction to grid middleware. The course also covers topics from current research in development and use of modern systems for distributed computing, including the use of cloud resources for grid computing. Virtualization is covered as a method to obtain task distribution on a global scale.

Learning Outcome

Upon completion of the course the candidate should be able to:

Knowledge

• discuss challenges and solutions for high volume data processing.
• explain the philosophy of cloud and grid computing.
• identify tasks well suited for execution on grid and cloud systems.
• assess selected research papers in the field of high volume data processing.
• explain the different cloud service models.
• describe the different hypervisor models used for virtualization.

Skills

• install and configure a representative grid middleware system (such as Globus).
• define and monitor job management, storage management and security in a grid system.
• design and implement applications of Service Oriented Computing at a global scale.

General competence

• evaluate and apply grid computing resources using textual and graphical interfaces.
• revise application software to make it suitable for grid and cloud execution.

Entry requirements

General admission criteria for the PhD programme.

Teaching methods

There will be 2-4 smaller exercises that must be approved in order to enroll of the exam. In addition, there will be a larger project covering problems to be solved using modern systems for distributed computing, cloud computing or virtualization. The project should include both a theoretic study and a practical problem solution. The theoretic study should be presented as a lecture and the prac-tical solution in a shorter oral presentation. The project should also be documented in a written report, covering both the theoretic study and the practical problem solution.

Compulsory learning activities

There will be 2-4 smaller exercises that must be approved in order to enroll of the exam.

Assessment

Grading according to the A-F scale based on an oral exam and the project report. The project report will have a weight of 40 % in the final grade.

Course reductions

• DAT351 - Skyløsninger og Distribuerte Dataressurser for Høg-Volum Dataprosessering - Reduction: 10 studypoints