Découvrez comment Nuxeo utilise MongoDB pour permettre à sa plateforme de gestion de contenu de s’adapter progressivement à la charge et au volume. Nous présenterons notamment le processus et les résultats de notre dernier test de performance réalisé avec succès sur plus de 11 milliards de documents en utilisant MongoDB Atlas.
2. Agenda
Nuxeo & MongoDB
● What Nuxeo Platform provides
● How we use MongoDB
Benchmark
Nuxeo & MongoDB at Scale
● How we tested
● Performances and Scalability
with Nuxeo and MongoDB Atlas
Context Scalability
11. Context | MongoDB Nuxeo
Performances
No impedance issue
Simplify architecture
Bulk import throughput
Document level locking
Sharding support
Scale out
Distributed Architecture
Scale-out Read
Scale-out Write
Simple Data Model
No ORM
Easy access to JSON documents
12. Benchmark | Principles
From 1B, 2B, 3B to 11B documents
Scale infrastructure step by step
Test on production ready environment
Nuxeo Cloud: AWS MongoDB Atlas
Encryption and backup are on
Make it real
Full content model with document types and full-text
Generate consistent and clean data
Build a realistic content application
13. Benchmark | Challenges
Import 11B documents
● Leverage NVMe to maximize IOPS
● Scale Up to reduce memory pressure
● Scale Out / Sharding to spread the load
Cost Efficiency
● Tiered Storage / Application-level sharding
● Scale Up for Import / Down for normal operations
● Leverage MongoDB Atlas Elasticity / Transparency
17. Scale from 0 to 11B
• Scale up / out progressively
M40 => M60 => M80 => M200
• Maintain good performances
Leverage Atlas to transparently adjust to needs
• Scale up for bulk import
• Scale up for bulk reindex
Cost efficiency
• Adjust storage cost to SLA
Archives M80 ⇔ M40
11B docs Benchmark | Takeaways