POLARDB Meets Computational Storage - Efficiently Support Analytical Workloads in Cloud
Offered By: USENIX via YouTube
Course Description
Overview
Explore a conference talk from FAST '20 that delves into the deployment of computational storage drives in Alibaba Cloud to enhance analytical workloads in cloud-native relational databases. Learn about POLARDB's compute-storage decoupled architecture and the challenges of maintaining cost-effectiveness while pushing data-intensive tasks to storage nodes. Discover how computational storage drives offer a solution by leveraging in-storage computing power for efficient table scans. Examine the implementation challenges and design techniques developed to integrate this technology into Alibaba's cloud infrastructure. Gain insights into the first real-world deployment of cloud-native databases with computational storage drives, covering topics such as data path transparent compression, system evaluation, and unit testing on storage nodes.
Syllabus
Intro
POLARDB: Cloud-Native Relational Database
Table Scan in Storage Nodes
Computational Storage: A Very Simple Idea
CSD with Data Path Transparent Compression
POLARDB Meets Computational Storage
Computational Storage Drive Implementation
Implementation & Evaluation
System Evaluation
Conclusion
Unit Test on Storage Node
Taught by
USENIX
Related Courses
Understanding the Robustness of SSDs under Power FaultUSENIX via YouTube BetrFS - A Right-Optimized Write-Optimized File System
USENIX via YouTube F2FS - A New File System for Flash Storage
USENIX via YouTube DNA Data Storage and Near-Molecule Processing for the Yottabyte Era
USENIX via YouTube FAST '21 Work-in-Progress Reports
USENIX via YouTube