Why Cloud-Native Storage Is Your Biggest Latency Surprise

Episode 37 of Cloud Computing with Fexingo: Lucas and Luna dig into a problem that's catching engineering teams off guard in 2026 — cloud-native storage performance. They break down the difference between local SSD, network-attached block storage, and object storage in AWS, Azure, and GCP, using real numbers: the 1–5 millisecond baseline for EBS gp3 versus the 200–600 millisecond p99 of S3 GET requests. Lucas explains why Kubernetes stateful workloads often hit unexpected latency because CSI drivers add overhead, and Luna shares a case from a media processing startup that saw transcoding jobs double in duration after moving to cloud-native storage without re-architecting. They explore how NVMeoF, local SSDs, and new tiered cache layers are changing the game, and why your cloud bill might hide performance regressions. No generic advice — just specific trade-offs for containers, databases, and high-throughput pipelines. #CloudComputing #CloudNative #Storage #Latency #AWS #Azure #GCP #Kubernetes #NVMeoF #EBS #S3 #PerformanceEngineering #StatefulWorkloads #Technology #CloudArchitecture #FexingoBusiness #BusinessPodcast #DataStorage Keep every episode free: buymeacoffee.com/fexingo