Please log in to watch this conference skillscast.
As the technology community adopts more scale-out applications consisting of containerized services, operations teams are feeling more pressure in achieving quality without excessive spending on infrastructure. Although container orchestrators (Kubernetes, Mesos, etc) allows multiple containers on shared hosts, co-locating different workloads on a shared cluster may cause severe resource interference among the co-hosted services and as a result service level violations for those applications with stringent quality of service (QoS) requirements. These challenges apply to both long-running services as well as analytical batch jobs. In this talk, Timothy will propose a strategy that combines an in-depth profiling and analysis of various interference types and their impact of application QoS and a semi-automated interference mitigation process for co-located application services.
YOU MAY ALSO LIKE: