Simulation Modelling and Analysis of Computer Networks Assignment 1

ASSIGNMENT # 01

I will post all my work here for FREE , hoping it would be beneficial for someone in future

If you find it helpful in anyway and want to thank, you can always

donate me on my bitcoin  16moPv4pyFX8etZxiiVAderXFo7pPaH3JE

:

like/share my page  my page www.kashifislamview.blogspot.com

TOTAL MARKS

Question No. 01 (20)

17. From published literature, select an article or a report that presents results of a performance evaluation study. Make a list of good and bad points of the study. What would you do different if you were asked to repeat the study

• Suggestion: Each student should select a different system such as a network, database, processor, and so on
  
  
  
  
  Performance evaluation of cloud application
  With constant data center configuration and
  Variable service broker policy
  Using CloudSim
  Kashif Islam
  Postgraduate Student Dept. Computer Engineering
  CASE, Pakistan
  kashif.islam.se@gmail.com
  
  
  
  
  Abstract
  Say you are starting a business, a small software company. Lets say 100 employees. Each employee should be provided a separate laptop/desktop. Operating System , along with costly applications should be installed on each system. Your proprietary software should be installed on each system. Now when upgrading , you have to upgrade 100 computers. There will be additional cost of upgrading hardware, faulty equipment replacement ,IT infrastructure etc. Cloud computing presents you with an alternative solution.
  Platfrom as a Service (PaaS), Software as a Service (SaaS) and Infrastructure as a service (IaaS).i.e. instead of buying 100 high performance costly desktop Pcs/laptops. You buy 100 cheap thin clients of netbooks. All your software including OS , will be on the cloud.Microsoft applications , your proprietary software will be in cloud. Now for upgrades you can schedule without any downtime.IT infrastructure cost will be reduced. You will only pay for the time employees are logged in and working. When people are not working they won't be utilizing resources on the cloud hence you won't be charged for this time.Now consider a second example where you build a social networking site. Users from every part of world log in to your side ,upload pictures, movies etc.
  At certain time different number of users logged in are different compared to other. In different Geographical locations. Let’s say at time 00:00 you have 1 million users utilizing system resources of cloud.At peak times you have 30 million users utilizing system resources. Power of cloud is it will expand resources for peak time and would reduce resourced during idle or less traffic thus fruitful for both customer and service provider. Now you know about your company timing, your load, peak load hours, complexity of application etc. Before leasing any cloud service , you want to evaluate performance of your application on cloud.
  There are different softwares available for simulating this. GridSim, CloudSim, GangSim, Micro Grid etc.
  GridSim actually simulates cost and performance for applications distributed applications deployed on cloud computing systems.
  While CloudSim is built on top of Grid Sim softwares is to the date best software for studying cloud computing infrastructures.
  Cloud Analyst built on top of  CloudSim  provides simulating controlled data centre configuration performance based of different metrics.
  
  
  Good points of study
  Well Defined Goal:  Goal of study was well defined to evaluate performance of cloud based on different policies for a cloud deployment of social networking application.
  
  
  Best Evaluation Technique: Best Evaluation technique currently present is used to simulate the problem.
  Social Aspects Included: Social aspects are included very thoroughly as users from different geographical regions will encounter different latency depending upon the performance metrics selected
  Realistic Workload: A realistic workload is selected in terms of number of users and geographical representation of load is showed in performance calculations.
  
  
  Presentation Style: Presentation Style used is suitable even for layman, while briefly including technical details with simplicity.
  
  
  Bad points of study
  Statistical in-correction: As a matter of fact huge social networking applications make use of customized SQL database on non-cloud platforms. Use of cloud technology is relatively new and migration costs are higher. Cloud is more effective for small scale applications which grow over time.
  
  
  Failing scenarios absent: Study does not include any failure scenarios, e.g hardware failure, network failure, application bugs, infinite loops.
  
  
  Security Concern absent: Cloud applications pose a security concern among many customers. Due to this reason customers are reluctant to migrate their Infrastructure to cloud. This paper does not evaluates any security concern with performance measuring. Normally implementing different security mesaures have considerable overhead on performance in terms of network and processing delay.
  
  
  
  
  
  What would I do different
  
  
  Include Application complexity factor. Author has not included application complexity factor in his research. Depending upon application complexity, latency can play a huge role in cloud apps.
  
  
  Include Customized number of Threads in Evaluation Techniques: If you are playing a heavy game , e.g let’s say Strong Hold Crusader on a cloud platform. Latency of some factors might be of less importance compared to others.
  e.g map navigation should become most efficient in terms of latency compared to selection of units.
  
  
  Better analysis and interpretation of data: If I would do research I would actually include analysis of different services offered  by the platform which are interdependent  on other cloud apps and represent data in terms of relational model.
  
  
  REFERENCES
  
  
• Android Anti-forensics: Modifying CyanogenMod  http://arxiv.org/abs/1401.6444 , Jan 2014

Performance evaluation of cloud application with constant data center configuration and variable service broker policy using CloudSim  International Journal of Enhanced Research in Science Technology & Engineering, ISSN: 2319-7463

Vol. 3 Issue 1, January-2014, pp: (1-5), Impact Factor: 1.252, Available online at: www.erpublications.com

I will post all my work here for FREE , hoping it would be beneficial for someone in future

If you find it helpful in anyway and want to thank, you can always

donate me on my bitcoin  16moPv4pyFX8etZxiiVAderXFo7pPaH3JE

:

like/share my page  my page www.kashifislamview.blogspot.com