KubeCon + CloudNativeCon India 2024

Vector databases like Milvus and LanceDB are revolutionizing similarity search and AI workloads. However, their performance in cloud-native environments depends heavily on optimized storage configurations. This session will delve into configuring Ceph's RADOS Gateway (RGW) using Rook for this workload. And provide a sample demo of how to run these applications with RGW.

In this comprehensive session, we will delve into the intricate details of data protection in the context of cloud-native applications. We will start by examining the automation of application discovery. This includes a special focus on application-consistent backups, particularly crucial for distributed applications utilizing multiple persistent volumes. We will then navigate the considerations for temporary storage and compute needs during backup and restore operations. The discussion will extend to the impact on production applications' compute and I/O performance. As a highlight, the session will explore strategies for detecting and protecting applications from ransomware attacks. The final segment of the presentation will cover application recovery and thereby mobility of applications across different Kubernetes platform distributions. Each aspect is designed to prepare participants for a future where cloud-native data protection is efficient, resilient, and cost-effective.

In the telecom industry, service availability is critical, making local and geographic redundancy essential for disaster recovery. While 5G Core Networks emphasize centralized, cloud-native strategies, 5G Access Networks require distributed, low-latency redundancy for edge cloud environments. Current CNCF tools like Velero, Portworx, and Stash offer Backup and Recovery Solutions but lack live synchronization between active and standby clusters, crucial for telecom's KPI needs. This session will explore a K8S-native solution designed to bridge this gap, offering seamless redundancy for telecom workloads, particularly in microservices-based deployments like vRAN/ORAN. This solution supports a 2-way policy-based sync, allowing for fine-grained control on how data is managed during failover. The solution integrates with CNCF projects, including Nephio for intent-based automation and Prometheus for real-time monitoring, aligning with the move towards distributed cloud-native deployments.

Embark on a journey to supercharge your Go applications targeting WebAssembly by harnessing the profiling capabilities of pprof and wzprof. This session unravels the intricacies of optimizing Go-powered web applications for maximum performance. Discover how pprof provides deep insights into CPU and memory usage, forming the foundation of our optimization journey. Complementing pprof, wzprof, tailored for WebAssembly, offers streamlined performance analysis during module execution. Through practical demonstrations, learn how pprof and wzprof work together to resolve performance bottlenecks, optimize computations, and manage memory effectively. This talk equips both seasoned Go developers and WebAssembly newcomers with essential tools and techniques to maximize application efficiency and speed.

With user namespaces reaching beta in Kubernetes and new developments in CRI-O, we’re closer to making nested containers within pods more flexible and powerful. Traditionally limited by masked /proc and restricted user namespaces, this approach now offers capabilities similar to Podman. In this talk, we will explore how Kubernetes’ security features—privileged mode, rootless containers, and network isolation—can enable running containers inside pods. We’ll examine the support matrix for various configurations and discuss upcoming work to bring VM-like flexibility to Kubernetes pods for more secure and dynamic container orchestration.

Through the years, ClusterAPI has evolved into an indispensable tool, streamlining the lifecycle management of Kubernetes clusters across multiple infrastructure providers. The current approach adds a layer of complexity at the image-building stage, presenting users with a multitude of options. But what if we challenge this conventional approach? This presentation introduces a paradigm shift in ClusterAPI image building, leveraging systemd-sysext and image composability. Join me in this talk as we explore how this innovative approach could help cope with the never-ending matrix of Kubernetes versions and Distro images, significantly enhancing usability for users managing their workloads.