Case Study
Drupal Content Management System
Autoscaling QA Infra Supports Massive Workloads for the 2nd Largest Open-Source CMS in the World
Situation
A Slow On-premise System that was Expensive and Unable to Scale.
The Drupal Content Management System (CMS) is the 2nd most popular CMS and one of the largest open source projects in the world, powering ~10% of the Top 100,000 sites & ~3% of the Internet. Every code change is aggressively tested against many permutations (various databases, operating systems, etc.) to ensure bug-free code is delivered to the millions of websites that rely on it. The original on-premise system was expensive, slow, & scaled poorly, creating bottlenecks that inhibited the platform’s innovation and growth
Solution
Improving Architecture Performance
Transform the application by improving architecture and performance. Design, build & deploy an auto-scaling, cloud based infrastructure enabling innovation, growth & continuous advancements. .
Results
Improved Performance. Auto Scalability. Optimization and More.
Performance: Unlike on-premises, cloud instances are easily adjustable yielding faster per-server run-time performance. Auto scalability: 100 instances at peak load, each with 32 cores running 30,000~ QA test assertions an hour. Over 7x faster. High-performance Computing: Concurrently processes 10+ years of compute time (testing) in each calendar year. Task Parallelization: Parallelized existing functionality and created an application performance testing framework, funded by Google, to track and improve CMS and QA performance. Optimization: System logs & app runtime data identify bottlenecks; tuning expedites results & lowers server costs.