Browsing by Author "Ataş, Gültekin"

Filter results by typing the first few letters
Now showing 1 - 1 of 1
  • No Thumbnail Available
    Item
    Performance evaluations of cloud computing platforms
    (Bahçeşehir Üniversitesi Fen Bilimleri Enstitüsü, 2013-06) Ataş, Gültekin; Güngör, Vehbi Çağrı
    There are many different types of cloud computing offers in the information technologies market. This study deals with the performance evaluations of three public “platform as a service” solutions offered by different companies. Cloud Foundry, Heroku, and OpenShift are the selected platforms for comparison. The market researches show that the cloud computing services will be one of the most demanded services that companies will look for in near future. Thus, companies will want to know which solution is good at which function. Similar to the other solutions, there is no silver bullet in cloud computing. Hence, the comparison must be based on the common things and those things should be the most used or needed resources. Therefore, computing power, database operations, and the main memory bandwidth are tested and compared in this study. The test results show that each platform performs differently in each of the area tested. A provider may be the best in one of the functions or sub functions of the performance criteria; however, it may be worse or the worst in another. This thesis proposes the most critical functions and sub functions for a PaaS solution. It also proposes a suitable set of benchmarking algorithms and suggests the best provider based on the test results by using two different decision making methods, and by taking the needs of the customer into account. Since the evaluation is based on the performance, the fees are not compared and not considered as a criterion. Similarly, other factors, like user friendliness, supported add-ons and services are nice to have; nevertheless, they are not the basic and common drivers of company needs in general. The performance results of the three platforms are compared to each other by using Analytic Hierarchy Process (AHP) and Logic Scoring of Preference (LSP). Four different scenarios are presented and the evaluation results and rankings of AHP and LSP are provided for each.