Gyu Tae Bae
Software Engineer, Container Platform @ NAVER Corp | CNCF Speaker & Open Source Contributor
Seoul, South Korea
Actions
Gyutae Bae is a software engineer on NAVER Corp.’s Container Platform team. He works on large-scale Kubernetes networking with Cilium/eBPF, focusing on reliability and performance across at scale. He diagnosed and fixed a connection-stability issue in Cilium and contributed the patch upstream. At KubeCon + CloudNativeCon North America 2025 (previously a KubeCon 2024 speaker), he shares pragmatic, production-tested practices—how to reason about BPF map pressure, avoid TCP resets, and turn incident learnings into platform tooling that benefits everyone.
Taming the BPF LRU: Eliminating TCP Resets in Cilium
At scale, Cilium users often face mysterious TCP connection failures from unexpected RST packets. This session explores a critical bug where Cilium's BPF-based SNAT and its LRU eviction policy prematurely terminate active sessions. We will dissect the root cause in the eBPF datapath and reveal the elegant fix, now merged upstream in Pull Request #37747: proactively restoring the original NAT entry on the reverse traffic path. This solution, born from a real-world production issue, reduced connection failures from up to 10% to nearly zero.
This talk is a must for operators debugging network instability and developers tackling real-world eBPF challenges. You will leave with a clear diagnosis for this "silent killer" and key insights into building robust, high-performance cloud networking.
Architecting Resilience: Lessons from Managing 7K+ Kubernetes Clusters at Scale
As a Kakao’s private Kubernetes as a Service team member, we manage over 7K+ clusters, 100K+ nodes. Due to a data center fire that occurred last year, we experienced significant economic and social impacts.
Many developers within the company utilize Kubernetes clusters, and the various services which run on Kakao's services each use multiple clusters. In this situation, failure in a data center would affect multiple services.
Therefore, cluster high-availability has become an important consideration, and we have been thinking about how to provide highly available Kubernetes clusters more efficiently for developers. In this talk, we will describe the design ideas we had for providing highly available Kubernetes clusters and the various problems and concerns we encountered while implementing them.
CNCF-hosted Co-located Events North America 2025 Sessionize Event
KubeCon + CloudNativeCon Europe 2024 Sessionize Event
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