Multi-Cloud Simulation

Extending CloudSim to enable support for multi-cloud simulations

Today, cloud based services are an integral part of our everyday lives. We use Dropbox, Google Drive and OneDrive to store files; Google Docs and Office 365 to write documents; Gmail, Yahoo mail and Outlook.com to write emails; and we develop and run our own web services and applications using Google App Engine, Amazon Web Services, Windows Azure Cloud Services – to mention just a few cloud services and cloud service providers. Traditionally, cloud applications are deployed on the infrastructure provided by a single cloud provider, imposing vendor lock-in, challenging application migration, and lack of flexibility. To meet these challenges, multi-clouds and intercloud infrastructures are fast becoming an emerging extension to the current cloud computing ecosystem. In a multi-cloud environment, applications are deployed in such a way that different parts of the application may span multiple public or private clouds simultaneously.

A multi-cloud app utilizing Amazon AWS, Rackspace, and a distributed private cloud!

Description of Topic:

As like any other emerging computational domain, cloud computing is also full of unsolved research challenges. To make clouds more efficient, challenges such as achieving workload optimization, higher predictability of services, and energy efficiency, to name a few, must be tackled. This often requires designing new algorithms, methods, and technologies. However, as the real strength of the proposed methods and algorithms may only be shown on a very large scale infrastructure with hundreds of thousands of virtual machines in place, it is extremely costly to conduct repeated experiments for evaluation. Thus, large-scale modelling and simulation is vital for the evaluation of new research in the domain of cloud computing. While there are several good tools, such as CloudSim [1], for the modelling and simulation of workloads running on a single-provider cloud systems, workload simulations on multi-clouds are not well supported. In this thesis, we target multi-cloud workload simulation and aim to design and develop a simulation framework for multi-clouds.

 

Goal:

The goal of this master project is to extend CloudSim to enable workload simulations on multi-cloud infrastructures.

 

What you will do:

  • You will extend CloudSim for multi-cloud execution.

  • You will analyze and evaluate different algorithms for workload distribution among multiple cloud domains.

  • Potentially publish your results at a good scientific venue!

What you will learn:

  • You will get to know state-of-the-art cloud computing infrastructures, topologies, and services.

  • You will develop a thorough understanding of workload execution in a cloud computing data center.

  • You will learn techniques and tools for large-scale simulations.

  • You will learn performance analysis and develop insight into algorithm evaluation.

Qualifications:

  • First-class programming skills in Java (or C/C++).

  • Previous experience or knowledge of cloud computing is a plus.

 

For more information, please contact Feroz Zahid (feroz@simula.no) or Ernst Gunnar Gran (ernstgr@simula.no).

 

References

[1] CloudSim: A Framework for modeling and simulations of cloud computing infrastructure and services

Emneord: cloud computing, multi-cloud, cloud simulation
Publisert 9. sep. 2016 11:48 - Sist endret 9. sep. 2016 15:11

Omfang (studiepoeng)

60