Take Your Percona Server for MongoDB to the Next Level
summary
Baseline RAID10 with snappy vs RAID5 plus Pliops XDP and built‑in line compression shows concrete wins: capacity rises from 9 TB → 12 TB on the same hardware, bulk load throughput jumps ~1.9x–2.3x, latency cuts (example 0.3 ms → 0.163 ms), disk throughput roughly doubles (169 MB/s → 353 MB/s), dump time drops 102 → 61 minutes, rebuild impact stays under ~10–15%, and CPU‑per‑core performance nearly 2x, all demonstrated with YCSB and real datasets.
AI-written summary · about
chapters (16)
- 0:00Intro
- 1:47Application Demands and IT Design Goals
- 4:41MongoDB Testing Environment
- 5:14Capacity Benefits with Pliops
- 7:13Data Loading Performance
- 8:08Performance: Data Loading Summary
- 9:15ETL, CDC & Message Queues Operation Benefits
- 14:21CPU Performance Comparison
- 16:22MongoDB Backup with Mongodump
- 21:21Performance during SSD Failure & Rebuild Operations
- 25:47MySQL Demo - SSD Write Amp Reduction
- 27:33MongoDB Scaleout Setup
- 33:36Storage Engine Performance Boost with Pliops
- 34:17Application 10 Amplification Challenge
- 35:38Pliops Key Value Technology Stack Higher Customer Value
- 37:22MongoDB Server Sprawl Reduction
related talks
24mPercona Server for MySQL Advantage
25mPercona Server for MongoDB: Features Overview
25mThe future of databases: Combining operational and analytical data at massive scale
51m5 boring performance improvements in MongoDB 3.4 that matter
09mOpen Software for MongoDB Where Percona Fits
37mThe State of MongoDB, Its Open Source Community, and Where Percona Is Going With It
39mPhysical Backups with Percona Backup for MongoDB
11h 21mPercona Community Live Conference - Day 3 - Open Source Databases Conference