For Students

Available theses

IS green metrics for Cloud infrastructures

Energy efficiency in Cloud infrastructures needs to be properly measure and evaluated. As a consequence a set of energy-related metrics has to be defined allowing for the determination of measurable energy-efficiency criteria at various level. These metrics will allow comparison of the energy efficiency and CO2 footprint of Cloud infrastructures and applications at various level, enabling to determine the most suitable infrastructure or mode of deployment for a specific application/service.

Goal of this theses is to analyze the state of the art on Green IS and to propose a set of measurable metrics at different level (IaaS, PaaS, SaaS). Relationships and dependencies among these metrics will be defined as well. A private cloud infrastructure will be made available to validate the defined metrics.

Contacts:

  • Cinzia Cappiello (cappiell@elet.polimi.it)
  • Pierluigi Plebani (plebani@elet.polimi.it)

Application energy consumption patterns

Energy Savings in Cloud infrastructure can be improved by properly tuning the running applications. Instead of analyzing and possibly modifying the internal structure of these applications, the idea is to analyze and compare the application energy profiles, that will include the trends of the energy-related metrics over the execution time.
Goal of the thesis is to define the application profile model and to propose an approach for analyzing such profiles in order to figure out the best practices and how to apply these best practices for other applications

Contacts:

  • Cinzia Cappiello (cappiell@elet.polimi.it)
  • Pierluigi Plebani (plebani@elet.polimi.it)

Co-opetitive application adaptation in Energy-aware Cloud infrastructure

Adaptation in Cloud infrastructure might be required at both application level and infrastructural level. About the former case, the goal is to optimize the performances (also including the energy-related metrics) of the application. On the contrary, in the latter case the objective is to find a global optimum. For this reason, the selection of adaptation actions at the two levels might clash.
The goal of the thesis is to define and develop an approach for managing this situation finding a solution that possibly bring to a win-win scenario

Contacts:

  • Cinzia Cappiello (cappiell@elet.polimi.it)
  • Pierluigi Plebani (plebani@elet.polimi.it)

Sensor as a service retrieval

Sensors are now deployed everywhere and they are able to monitor and measure several physical quantities. Recently, sensor networks have been joined the Service Oriented Computing paradigm. As a consequence an abstraction layer composed by several services is available to communicate with the underlying sensors in a transparent way.
Goal of the theses is to propose a service retrieval algorithm. Assuming that a huge number of services are available and indexed in a registry, the algorithm will support the user when measures of a physical quantity is required in a given context. The theses will combine competencies on Service Oriented Computing and Semantic Web.

Contacts:

  • Pierluigi Plebani (plebani@elet.polimi.it)

Contingency plan similarity

Emergency preparedness is considered the first phase of emergency management, followed by response, recovery and mitigation and refers mainly to emergency planning, training and exercising aspects. During this phase the emergency planning process aims to prepare all concerned organizations to face a given crisis when a disaster occurs due to an accident or a natural extreme. This process results in a document called contingency plan.
Goal of the theses is to define a contingency plan management system that provides mechanisms to define and select contingency plan relevant in case of earthquakes. A crucial element of this system is the contingency plan similarity algorithm that allows to find contingency plans closer to the emergencies that is occurring.

Contacts:

  • Pierluigi Plebani (plebani@elet.polimi.it)