The schedule of tutorials is shown below.
| Tutorial 1 | ||||
| Title | ||||
| Speaker1 | ||||
| Speaker2 | ||||
| Speaker3 | ||||
| Duration | ||||
| Abstract | This tutorial aims to introduce the audience to new automated methods of performing cloud analytics for improving the efficiency, security, and resilience of cloud deployments. While machine learning (ML) has proved useful in a wide variety of modern computing applications, today’s cloud systems still rely on the manual efforts of expert human administrators and ad-hoc tools. In this tutorial, we share a new perspective on an area of cloud administration known as system discovery, which focuses first on building representations of systems and software that are easy to parse and organize, and second on designing learning-based frameworks to identify properties of interest. This tutorial focuses on software discovery, where the goal is to identify the software contents of large-scale cloud deployments that include short-lifecycle containers and VMs. Talks feature experts from the cloud analytics field covering real-world scenarios where ML-based methods achieve significantly better outcomes than previous approaches. Participants can experience these benefits directly during handson activities through the use of open-source analytics frameworks designed by the speakers’ teams at IBM and Boston University. Participants will leave this tutorial with a new awareness of ML-based cloud operations and the experience necessary to use ML-based tools in the field. | |||
| Tutorial 2 | ||||
| Title | ||||
| Speaker1 | ||||
| Speaker2 | ||||
| Duration | ||||
| Abstract | Kubernetes has become the industrial go-to platform for composite microservice workloads forming cloud-native applications. Yet it is surprisingly low-level and not easy to start with from a software engineering perspective. In this half-day hands-on tutorial, we briefly introduce the platform and the conventional approach to develop Kubernetes applications, before diving into two relevant research topics with recent results: Quality assessment of such applications, and feedback loops based on the Kubernetes operator framework. On the quality side, we look jointly at packaging solutions including Helm and Kustomize which have gained traction with developers but are not yet on the radar of most researchers. On the feedback loop side, we specifically dive into annotations on deployment-related objects, and report on our experience on automating monitoring and incident management on this basis. We explain how the combined approach, to quality-assess annotations during the development process in continuous development pipelines, can increase the quality of self-managed and cloud-native applications. | |||
