Side Channel Analysis and particularly the power analysis has been an open problem since it was first discovered. As of today, the world is moving towards the Internet of Things (IoT) and hence the devices are vulnerable to these attacks. The countermeasures are guarding, masking, and rekeying. Most of the countermeasures the researchers are proposing are related to masking and rekeying. However, their approach is generic algorithms due to various existing and new devices that come up, and their comparative analysis focuses mostly on mathematical leakage models. The devices are also ever evolving with newer versions updates. Each different version of the device would contain different vulnerability. Therefore, the algorithms are not assessed on the implementation aspects. SILK simulator [5] is a useful tool to simulate the traces for the various Arduino family devices. They are configurable for the particular device and can be powerful tool to validate the proposed methods in the implementation scenario that would provide accurate effect and assessment of the approaches. It is also required to validate the simulations with the actual hardware devices as the actual environment would add its additional noise. In this paper, the authors have validated their proposed and published algorithm [1] on the SILK simulator and on the actual Arduino Uno device. The aim of the study is to check the actual vulnerability of the simulation and the actual device and compare them along with the assessment of earlier algorithm on the actual hardware environment. A total of 10000 power traces were generated in simulator and by the actual Arduino Uno device for 40 different keys for both the plain AES and modified AES based on our previous work to check the time and efforts required to successfully recover the key using the Correlation Power Analysis (CPA) attack. The results show that our proposed published method has shown more resistance to the attack compared to the AES with difficult on the actual device compared to simulator. However, the simulators provide a better picture for algorithmic comparison than the mathematical model due to device specific configurations.