Kahiro Okina
Advocate for collaboration and innovation in the tech community.
テックコミュニティでのコラボレーションとイノベーションを推進するアドボケイト
Tokyo, Japan
Actions
Tech Lead at Craftsman Software, Inc. and #8 CNCF contributor in Japan. Active in the Kubernetes and CNCF ecosystem as a maintainer of the Headlamp Knative plugin, Cluster Inventory API, ClusterProfile Integration for Argo CD, and Open Cluster Management, and as a reviewer for Knative Gateway API, Knative Operator, and Cluster API Provider k0s. He also serves as an instructor for Kubernetes Upstream Training Japan and as a mentor in the LFX Mentorship program.
Craftsman Software, Inc. のテックリードであり、日本における CNCF コントリビューター第8位。Kubernetes / CNCF エコシステムでは、Headlamp Knative プラグイン、Cluster Inventory API、ClusterProfile Integration for Argo CD、Open Cluster Management のメンテナーとして活動し、Knative Gateway API、Knative Operator、Cluster API Provider k0s のレビュアーも務めている。また、Kubernetes Upstream Training Japan の講師を務め、LFX Mentorship プログラムではメンターとしても活動している。
Area of Expertise
Zero-Touch Multi-Cluster Onboarding with Cluster API & Cluster Inventory en
Spinning up clusters is fast with Cluster API, yet wiring each one for production still drains time and patience. I’ll spin up—or register—a cluster, let Cluster Inventory auto-install Gateway API + Knative Serving, deploy a scale-to-zero service, then trigger a failure and watch it reappear on a standby cluster—no extra repos, no copy-paste YAML. Attendees will see a repeatable pattern that slashes setup toil, boosts resilience, and keeps idle-compute spend near zero.
From Zero to Production: Instant Multi‑Cluster Rollouts with Cluster API & Cluster Inventory en
Platform teams face a growing challenge: every project demands its own Kubernetes cluster, yet each cluster requires identical production-grade setup. While Cluster API streamlines cluster provisioning, the day-2 operations remain painful—teams resort to copy-pasted manifests, per-cluster repositories, cluster-specific configurations, and increasingly unmaintainable deployment pipelines.
This session demonstrates a practical solution to these operational challenges. Through live demonstration, I'll show how Cluster API and Cluster Inventory work together to transform multi-cluster management. You'll see automated addon deployment, scale-to-zero workload management with Knative, and automatic failover between clusters—all without duplicating configurations or scripts.
Key takeaways: Learn to provision clusters on-demand, automate addon and application deployment across clusters, reduce idle compute costs through scale-to-zero architecture, and ensure service continuity with automatic cluster failover. This approach dramatically reduces onboarding time while improving both availability and cost efficiency.
KEP-1645 Compliant Implementation Based on kube-proxy and Integration with Envoy Gateway en ja
This session introduces a specific method for implementing the MCS-API, whose official implementation method is not determined, using only kube-proxy while complying with KEP-1645. First, we'll explain the mechanism of MCS-API and how to establish service connectivity between multiple clusters simply and reliably using kube-proxy, detailing the implementation flow. Then, through a working demo implementation, we'll deep dive into the key points of building and operating Multi-Cluster Services. Furthermore, we'll introduce an approach to integrate MCS-API with Gateway API using Envoy Gateway. We'll provide detailed explanations of the necessary configurations and processing flows for achieving more advanced and flexible traffic management between multiple clusters, using architectural diagrams and concrete examples. By combining MCS-API and Gateway API, you'll gain insights into the advantages and optimization tips for scalably publishing and operating services.
KEP-1645に準拠したkube-proxyベースの実装とEnvoy Gatewayとの統合 en ja
本セッションでは、公式に実装方法が定まっていないMCS-API (Multi-Cluster Services API) を、KEP-1645に準拠しつつkube-proxyだけで動作させるための具体的な方法を紹介します。
まずはMCS-APIの仕組みと、kube-proxyを用いていかにシンプルかつ確実にマルチクラスタ間でサービスを連携させるのか、その実装の流れを解説します。
そのうえで、実際に稼働するデモ実装を通じて、Multi-Cluster Servicesの構築・運用のポイントを深掘りします。
さらにEnvoy Gatewayを活用し、MCS-APIとGatewayAPIを連携させるアプローチを紹介します。
複数クラスター間でのトラフィック管理を、より高度かつ柔軟に実現するために必要な設定や処理の流れを、アーキテクチャ図や具体例を交えながら詳説します。
MCS-APIとGatewayAPIの組み合わせを適用することで得られるマルチクラスタ公開・運用の最適化のヒントを得られるでしょう。
KubeCon + CloudNativeCon Japan 2026 Sessionize Event
Kahiro Okina
Advocate for collaboration and innovation in the tech community.
Tokyo, Japan
Actions
Please note that Sessionize is not responsible for the accuracy or validity of the data provided by speakers. If you suspect this profile to be fake or spam, please let us know.
Jump to top