Exploring the Impact of Cloud Computing and Edge Computing on Resource Consumption for Mobile Devices with Generative Artificial Intelligence APIs

With the release of OpenAI’s ChatGPT, generative AI has become a hot topic in business. Generative AI APIs can be consumed by mobile devices, which are constrained by resources such as CPU, memory, and battery power. Power can be saved by using both Edge and Cloud Computing. However, there is a penalty to response time implementing these services. There seems to be a greater impact on response time with Cloud Computing than with Edge Computing, however Edge Computing is constrained by limits on payload and maximum response times. Our testing results show that Edge Computing is a good middle ground, providing a balance between savings in CPU time, while somewhat mitigating the penalty in response time. Developers exploring the option to use Cloud and/or Edge Computing for power savings on mobile devices should consider the size of their data sets carefully to determine if an Edge Computing service is a viable option.