1. CLOUD COMPUTING
BASED ON THE PAPER, “A SURVEY OF MATHEMATICAL
MODELS, SIMULATION APPROACHES AND TEST BEDS USED FOR
RESEARCH IN CLOUD COMPUTING”, BY GEORGIA SAKELLARI AND
GEORGE LOUKAS.
2. AGENDA
• What is cloud computing?
• Research in Cloud Computing
• Mathematical Modelling of Cloud Systems
• Cloud Simulation Software
• Cloud Test beds
• Commercial Cloud services used in research
• Research Cloud test beds
• Software frameworks for setting up private cloud frameworks
• Conclusion
3. WHAT IS CLOUD COMPUTING?
A BRIEF INTRODUCTION ON CLOUD COMPUTING
4. WHAT IS CLOUD COMPUTING?
• Cloud computing is Internet-based computing, whereby shared
resources, software, and information are provided to computers and
other devices on demand, like the electricity grid.
7. WHAT IS CLOUD COMPUTING?
• Cloud computing relies on sharing of resources to achieve coherence
and economies of scale similar to a utility (like the electricity grid) over
a network.
• The cloud also focuses on maximizing the effectiveness of the shared
resources. Cloud resources are usually not only shared by multiple
users but as well as dynamically re-allocated as per demand. This can
work for allocating resources to users in different time zones.
8. RESEARCH IN CLOUD COMPUTING
PERFORMANCE ANALYSIS, SIMULATION MODELS AND OTHER RESEARCHES CARRIED
OUT ON CLOUD COMPUTING
9. RESEARCH IN CLOUD COMPUTING
• Research on the energy efficiency and performance of cloud
computing primarily focuses on the IaaS paradigm.
• Other common cloud computing approaches include PaaS and SaaS.
• Validation of these models is done through simulation.
• But, majority of the researchers build their own, small scale, cloud test
beds.
11. MATHEMATICAL MODELLING OF CLOUD
SYSTEMS
• Mathematical modelling of cloud systems is less demanding on
hardware and software investment.
• It also provides an understanding about the interdependencies
involved in cloud computing.
• It is particularly suitable for identifying optimal values and equilibrium
and predicting behavior.
12. MATHEMATICAL MODELLING OF CLOUD
SYSTEMS
• Various researchers have worked on different mathematical models of
the cloud.
Mathematical Model Description of the focus
S. Islam et al Elastic characteristics of the cloud.
Abdelsalam et al Optimizing servers and frequencies to consume minimal energy.
Van et al Energy based utility maximization.
Mohsenian-Rad et al Energy information transmission tradeoffs.
Mi et al Virtual machine allocation modelling.
Baliga et al Energy consumption of a cloud.
Garg et al Energy characteristics of a cloud.
14. CLOUD SIMULATION SOFTWARE
• Grid computing simulators like GridSim, cannot sufficiently simulate
the cloud infrastructure.
• Virtualization can be used to deploy virtual private clouds on small
scale physical test beds.
• But for large scale simulations, we need something like, CloudSim
(based on GridSim).
• CloudSim can simulate up to 100000 machines in less than 5
minutes.
15. CLOUD SIMULATION SOFTWARE
Simulation
Software
Language Open
Source
Description
CloudSim Java Yes Modular and extensible framework able
to simulate very large clouds.
CDOSim Java No Enhances CloudSim by facilitating
integration of more fine grained models.
TeachCloud Java Yes Extends CloudSim. Used for
Educational purposes.
NetworkCloudSim Java No Extends CloudSim with an emphasis on
networks.
GreenCloud C++ Yes Simulates a cloud as a packet network
and estimates energy consumption.
17. COMMERCIAL CLOUD SERVICES USED IN
RESEARCH
• Amazon Elastic Compute Cloud (EC2) is the most used commercial
cloud service in academic research.
• Amazon EC2 provides IaaS at a very large scale.
• Researches are not stable with some indicating that Amazon EC2 as
not suitable for scientific research while others indicate it as a viable
candidate.
• More importantly, its architecture can be used for more dependable
simulations than EC2 itself.
18. COMMERCIAL CLOUD SERVICES USED IN
RESEARCH
Commercial Cloud Service Pricing
Amazon EC2 IaaS Commercial
Amazon S3 Storage Commercial
Google App Engine PaaS Commercial/Free
Google Apps SaaS Commercial/Free
Windows Azure PaaS/IaaS Commercial/Partially Free for
Students
20. RESEARCH CLOUD TEST BEDS
• OpenCirrus is a joint initiative by Hewlett-Packard, Intel and Yahoo! in
collaboration with more than eight other organizations and universities
around the world.
• It enables researchers with similar interests to exchange data sets
and develop standard cloud computing benchmarks.
• It uses the Hadoop Distributed File system
21. RESEARCH CLOUD TEST BEDS
Scientific Test bed Description
OpenCirrus Large scale cloud test bed spanning over 10
locations worldwide.
Open Cloud Test bed Large scale cloud test bed supporting computations
over more than one data center. (Not possible in
OpenCirrus)
Science Clouds Cloud research test bed able to connect to Amazon
EC2.
Virtual Computing Lab Private cloud providing free usage to academic
institutes for limited amount of time.
Grid 5000 Test bed for cloud based massive data analytics.
23. SOFTWARE FRAMEWORKS FOR SETTING
UP PRIVATE CLOUD FRAMEWORKS
Software
Framework
License Service Description
Eucalyptus Open Source IaaS Generalist open source framework running
primarily on Ubuntu and CentOS
OpenStack Open Source IaaS Platform with the most active community
and most industrial approval.
Abiquo Proprietary/Open
Source
IaaS Platform manager for private clouds. Open
Source version has limited functionality.
Tplatform Open Source PaaS Designed for Web mining applications.
mOSAIC Open Source PaaS/Iaa
S
Provides interoperability between different
clouds.
25. CONCLUSION
• Cloud computing offers a lot of scope for research.
• It is viable platform for creating, deploying and using applications.
• Commercial cloud frameworks do not provide the accuracy needed by
researchers.
• Amazon EC2 is the most used platform for research and analysis.
