Jeff Kim on Stateful Application Data and Kubernetes Integration at Edge

About Kmesh sells software that transforms your centralized data into distributed data, which operates over multiple clouds, countries and edges as a single global namespace. Customers configure their business rules / data orchestration policies through the Kmesh portal and easily deploy their data over any cloud (production data on prem, backup data on AWS, AI/ML data on Google, IoT data on the Edge etc.)


· Kubernetes Challenges for Stateful Data

· Managing Stateful Data across Multiple Edge Devices

· How Kmesh works with Luster ~ New Global Data Models

· Kubernetes Limitations with Storage


· 0 min 24 sec: Introduction of Guest

· 2 min 36 sec: Data and Distributed Storage in Kubernetes Story

o Kmesh operates at Data Layer not App Layer as Kubernetes does

o Kubernetes is not the be all end all

o Kmesh arriving in parallel with Kubernetes (2.5 years old) — started as Luster as a Service for HPC

· 7 min 13 sec: How does Data Storage and Container Mgmt Come Together?

o Kubernetes has no state ; Data for Apps often times requires state

o Storing data with Kubernetes is a challenge — location, security, etc.

o Example: Car on road sending real-time data to Edge which has it connecting to various edge devices which need to have data properly ordered, managed, etc

· 10 min 34 sec: How manage multiple edge devices and data state across them

o Luster is low-latency, high-ops technology

o Ported Luster into cloud and edge environments

o Do you put the data in the right spot to take advantage of data store?

o Data duplication issues? Metadata is the key

· 19 min 24 sec: What does Kubernetes need to do to take advantage of this technology?

o Kmesh offers customer plug-in for Kubernetes

o Storage advancements in Kubernetes is limited

o How make Kubernetes workload better understand data?

· 27 min 02 sec: How does SaaS play into distributed data?

o Control plane vs Data plane

· 29 min 31 sec: Wrap-Up

Leave a Reply