Towards a serverless computing and edge-intelligence architecture for the Personal-Internet-of-Things (PIoT)
Ankur Gupta, Surbhi Gupta · 2024
Serverless computing has emerged as a compelling paradigm for the deployment of applications at the edge facilitating performant and lightweight services. Services deployed at the edge typically consume lesser compute and storage resources while offering relatively lower latency. Hence, a serverless architecture is a natural fit for the requirements of the Personal-Internet-of-Things (PIoT), an emerging sub-domain of the IoT in which data acquisition, analysis and computation is performed in the context of an individual user and her environment. For such environments privacy preservation, resource-light models, fast detection of emergency situations and low response times for services is paramount. In addition users demand hyper-personalized service delivery requiring embedded intelligence for real-time interventions. This paper proposes a serverless computing architecture with embedded intelligence at the edge for the PIoT which effectively meets the stated requirements. A proof-of-concept using AWS Lambda and tiny AI is implemented and initial results presented, establishing the effectiveness of the proposed framework.