Case Study: CloudFlare achieves 95% fewer dropped packets and scalable, DDoS‑resilient performance with Pluribus Networks

Supercharging the Internet with CloudFlare

CloudFlare, a global web performance and security company expanding to 100+ new data centers, needed to scale its CDN while keeping low latency and strong protection against attacks. The team was seeing excessive packet drops and lacked visibility into congestion and DDoS sources because existing top-of-rack switches had shallow buffers and couldn’t handle port contention.

CloudFlare standardized on Pluribus Networks’ Freedom F64 server-switches and Netvisor OS, gaining deep-buffer queue management, flow-based traffic classes, always-on analytics and an open API for rapid responses. The new fabric reduced dropped packets by over 95%, eliminated microburst-induced packet drops and TCP retransmissions, improved DDoS detection, and cut administrative costs by roughly 50%, enabling reliable global growth.

Cloudflare

Matthew Prince

Co-Founder and CEO