Large-Scale Practice of Persistent Memory in Alibaba Cloud

Explore a conference talk on the large-scale implementation of persistent memory in Alibaba Cloud. Discover how Alibaba leverages persistent memory to optimize resource allocation and reduce costs for millions of databases and memory-intensive applications. Learn about their innovative combined system for PMEM resource optimization and capacity scheduling, widely used in Kubernetes platforms. Gain insights into how PMEM devices are utilized as Kubernetes volume objects with various configuration types. Understand the optimization of schedulers for optimal NUMA node matching with PMEM devices. Delve into the architecture of Pmap, node resource management, and CPU scheduling techniques. Witness a demo showcasing the benefits of this technology and explore future developments in personal memory solutions.

Introduction
Agenda
Why is persistence memory
Necessary of using Pmap
Architecture of Pmap
Node Resource Manager
Scheduling
CPU Scheduling
Part 1 Introduction
Processing Memory Stacks
Benefits
Demo
Personal Memory
Future Work