Abstract:

Performance uncertainty is a major deterrent to cloud adoption, with consequences for cost, revenue, and performance. Predictable performance becomes increasingly more crucial when cloud services transition from monolithic architectures to microservices. In microservices systems, discovering QoS breaches after they occur results in long recovery periods since spots propagate and amplify across dependent services. However, their usage in industrial datasets has been less frequent. Additionally, the logging statements in the open-source datasets being studied are often rather large and do not change much over time. For a brand-new dataset from an industrial service, this may not be the case. This study tests many state-of-the-art anomaly detection techniques on the industrial dataset from the project partner, which is much smaller and less structured than most large-scale open-source benchmark datasets. Therefore, microservices provide heterogeneity, scalability, agility, and a fair degree of fault tolerance by breaking up modular programs into several services. However, there are a number of issues and challenges with this architectural paradigm that businesses must deal with. The purpose of this article is to outline the primary advantages and possible drawbacks of the microservices architecture and to propose an intelligent microservices structure that leverages AI and ML to support automation, flexibility, and optimization. This is to ensure that there is a rationale for why a certain activity should be performed and how it may achieve the process's desired goals.

Description:

.