Workload forecasting in Cloud Computing environments

Motivation
Changes of the workload profile directly affect the usage of available resources. Therefore, resource configuration and reallocation techniques are applied to use these available resources efficiently. An essential element of proactive reconfiguration mechanisms in dynamic systems is a most accurate prediction of the workload change.

Goals
In this work, you investigate existing techniques for workload forecasting, e.g. techniques used in the financial markets like chart analysis etc. A summary of the advantages and drawbacks of these techniques and their applicability under different time horizons (minutes, hours, days) are another goal. In addition, another goal is the implementation of such a technique or the integration of existing tools in our framework for reconfiguration in virtualization environments. Finally, an evaluation with an industry-standard application should be conducted.

We offer you:

• Work with state-of-the-art and innovative technologies
• Closely related to current research projects
• Excellent working environment and intensive mentoring

We are happy to support sientific publication of excellent work.

Duration
6 months

Contact
Nikolaus Huber