Analysis of Cloud Computing Environment using CloudAnalyst
Abstract: The systems become extremely complex, large and expensive. Change is a must
nowadays. And companies want the deployment of change in extremely fast manner without
compromising the system availability. In recent years the Cloud computing has become very
popular and it introduces new trends in software engineering. The needs of cloud environment
are different from the others. It is used to deploy scalable applications in an efficient way,
allowing enterprise applications to dynamically adjust their computing resources on demand.
1. CASE STUDY
Here we explain how the simulation is setup and executed using CloudAnalyst.
Social networking is the use of Internet-based social media sites to stay connected with friends, family, colleagues,
customers, or clients. Facebook is a consumer network product, not technically SaaS, but there's no other product
that provides as many services as Facebook does.
A typical large scaled application type on the Internet today that could benefit from Cloud is
social networking applications. E.g. Facebook, one of the most popular social networking sites
has over 200 million registered users world wide. On 18/06/2009 the approximate distribution
of the Facebook user base across the globe is given in the table below[5]:
Facebook subscriber statistics as of January 31, 2020 is as follows [5]:
SETTING UP A SIMULATION [3]
We need to carry out the following steps.
1. Define user bases –
We defined the following user bases:
User Region Time zone Peak hours Peak hours Online users Online users
base Local GMT Peak hours Off peak hours
UB1 0 North America GMT-6 7.00–9.00 13:00-15:00 25280410 2528041
pm
UB2 1 South America GMT-4 7.00–9.00 15:00-17:00 41491890 4149189
pm
UB3 2 Europe GMT+1 7.00–9.00 20:00-22:00 39550611 3955061
pm
UB4 3 Asia GMT+6 7.00–9.00 1:00-3:00 83233640 8323364
pm
UB5 4 Africa GMT+2 7.00–9.00 21:00-23:00 21291170 2129117
pm
UB6 5 Oceania/Australia GMT+10 7.00–9.00 9:00-11:00 2305622 230562
pm
We consider some assumptions just to make our work simple here. The following assumptions are used here:
1. Here we assume that 2
2. 0% of registered users will be online in peak hours in a specified region.
3. Next assumption is that 10% of peak hour users will be online in off peak hours.
The following costing plan is used, which is close to Amazon EC2 costing plan.
Cost per Virtual Machine per hour = 0.1 $
Cost per 1 GB data transfer = 0.1 $
2. Define data centers –
Using the Data Centers tab of the Configuration screen define the data centers we wish to use in the simulation. We use Define
all the hardware and accounting aspects of the data centers here.
Two scenarios are used here:
Single data center is used
Multiple data centers are used
3. Allocate Virtual Machines for the application in Data Centers –
Depending on the Scenario used, we will allocate the virtual machines to data centers.
Executing Simulation
RESULTS
1. Using Single data center
2. Using Multiple data center
a. Two data centers with 30 VMs each
b. Two data centers with 50 VMs each
c. Three data centers with multiple VMs each
2. CONCLUSION
References
1. Reena Panwar, Bhawna Mallick, “Load Balancing in Cloud Computing Using Dynamic Load Management
Algorithm”, 2015 IEEE
2. Dr. Hemant S. Mahalle Prof. Parag R. Kaveri Dr.Vinay Chavan, “Load Balancing On Cloud Data Centres”, Volume 3,
Issue 1, January 2013, International Journal of Advanced Research in Computer Science and Software Engineering
3. Priyanka Rajput, Sanjeev Kumar, “Simulation of a Large Scaled Web Application on the Cloud Using Cloud Analyst”,
Volume 10, Issue 9 (September 2014), International Journal of Engineering Research and Development
4. Meeta Singh ,Poonam Nandal , Deepa Bura, “Comparative Analysis of Different Load Balancing Algorithm Using
Cloud Analyst”,
5. Facebook usage and Growth Statistics, Web Link: https://www.internetworldstats.com/facebook.htm
