Maximize database price performance with Aurora and Graviton2 instances

on YouTube ↗19m

" summary

Amazon Aurora separates compute and storage (auto-scaling to 128 TB, 4-of-6 quorum across AZs) and runs on AWS Graviton2 Arm cores that claim 20% performance and up to 35% price performance versus comparable Intel R5 instances. Benchmarks use HammerDB/TPCC, pgbench and sysbench showing R6g outperforming R5 (e.g., r6g.16xlarge > r5.24xlarge in some tests). Clusters support one primary + 15 readers and mix-and-match R6g/R5 instances for targeted cost/perf.

AI-written summary · about

" chapters (5)

1. 0:00Intro
2. 0:34Agenda
3. 1:10Why build Aurora
4. 5:42Why build Graviton2
5. 16:34Resources

" related talks

• 25mAWS Graviton: Discover the Power of Distributed SQL + GravitonHemant Bhanawat, Arthur Petitpierre · Distributed SQL Summit 2022
• 25mAdopt AWS Graviton2 instances for best price-performance in EC2Arthur Petitpierre · AWS AMER Summit May 2021
• 22mGravitonStephanie Shyu, Ali Saidi, Martin Yip · AWS Silicon Innovation Day 2023
• 55mAmazon Aurora: A deep-dive into what's newSailesh Krishnamurthy · Percona Live Open Source Database Conference 2017
• 43mDeep dive into Amazon Aurora and its innovationsOvidiu Hutuleac, Katja-Maja Kroedel · AWS Summit Berlin 2024
• 39mAWS AuroraBlair Layton · FOSSASIA Summit 2017
• 45mAWS Graviton: Best Price Performance for your AWS WorkloadsGuy Almog, Barak Nissim, Yotam Bagam · AWS Summit Tel Aviv 2023
• 49mBuilding real-time applications to utilize AWS GravitonSubbu Iyer · AWS re:Invent 2022