Case Study: Zillow achieves scalable, high-performance image delivery and lower costs with Amazon Web Services

Zillow - Customer Case Study

Zillow, a leading real-estate marketplace used daily by tens of millions to browse 110 million U.S. homes, was struggling to manage explosive image traffic—more than 3 million new images per day and peaks of about 17,000 image requests per second. Its legacy, data-center–based imaging pipeline (NAS-stored pyramidal TIFFs, local Squid caching, and a fragile queueing and processing toolset) couldn’t scale, suffered from feed-driven bottlenecks, and presented disaster-recovery risks.

Zillow migrated its imaging stack to AWS—using Amazon S3 (≈100 TB, ~300M images, >1B objects), EC2, Elastic Beanstalk workers, per-feed SQS queues, a throttled downloader, CloudFront, and CloudWatch—to decouple, scale, and monitor processing. The result: reliable delivery at roughly 15,000 images per second, improved availability and disaster recovery, faster deployments and troubleshooting, and substantially lower CDN and operating costs (CloudFront reduced CDN spend to less than half).

Zillow

Feroze Daud

Senior Software Development Engineer