A Microservice-based MLOps Platform for Efficient Development of AI Services in an Edge-Cloud Environment
Chorwon Kim, Geon-Yong Kim, Sungchang Kim · 2023
With the advancement of Internet of Things (IoT) and Artificial Intelligence (AI) technologies, outcomes derived from diverse sets of collected data are finding applications in the business domain. In particular, edge AI technology, which combines edge computing and artificial intelligence, emerged to address real-time response limitations, is garnering considerable attention as a solution to the real-time response constraints faced by AI services within operated cloud environments. Nevertheless, the development of AI services within an edge-cloud environment presents numerous challenges, such as limited computing resources at the edge, the portability of learning models, and dependencies on code libraries during the build phase. Consequently, various open source projects and companies are actively advancing MLOps solutions. In this paper, we describe the development of a microservice-based MLOps platform utilizing open source for the development of AI services and the deployment of generated models in an edge-cloud environment built using lightweight edge hardware, software, and designated servers assuming the role of a cloud system.