Maximizing Workload's Performance With Smarter Runtimes

The presentation discusses maximizing workload performance with smarter runtimes in the context of container runtimes and resource utilization.

• Mitigating noisy neighbors in container runtimes is a challenge
• Existing solutions have limitations
• New innovations in container runtimes can improve performance and resource utilization without modifying workloads
• The presentation covers existing extension points for containerd and CRI-O, new ideas from NRI proposal, and evolution of dynamic resource usage optimizations in CRI-Resource-Manager
• The importance of understanding how resources are used and measuring them is emphasized
• The presentation discusses the need for NRI to become the primary integration point for extending runtimes and enabling smart algorithms
• The need for declarative resource requirements, separation of what in the cubelet from the how in the runtime, and a dedicated API for container state monitoring is highlighted
• Flexibility and customization are important in hardware performance optimization

The speaker emphasizes that there is no one algorithm that can satisfy all hardware performance optimization needs due to the evolving nature of hardware. Flexibility and customization are key in achieving optimal performance.

Mitigating noisy neighbours in the world of containers is not an easy task. There are several solutions exists and many of those have own limitations. This presentation will be focusing on exploring new ways of innovations for container runtimes that helps get maximum performance and resource utilisation without modifications of the workloads. In this talk we are planning to briefly cover existing extension points for containerd and CRI-O, talk about new ideas from NRI proposal, as well as covering evolution of dynamic resource usage optimisations in our project CRI-Resource-Manager. We want to share our experience on dealing with heterogenous CPU resources, multi-tiered Memory, Caches, Memory Bandwidth and Block I/O usage. We want to demonstrate how using various metrics and hints provided by Linux kernel can lead to improvements of workload performance and dynamic hardware resource utilisation optimisations.