You appear to be offline. Some site functionality may not work.
Call Us

A new architecture for the modern internet

Deploying thousands of small, scattered points of presence (POPs) may have worked for legacy CDNs in the dial-up era, but the internet has become increasingly dynamic, and spinning disks no longer get the job done. Fastly has taken a fundamentally different approach: we’ve focused our efforts on placing fewer, more powerful POPs at strategic markets around the world. With Tier 1 transit, solid-state drive (SSD) powered servers, and an engineering team that lives to optimize for speed, we’ve built a blazing-fast network that requires less hardware to deliver comprehensive global reach. Fastly’s high-density POPs enable us to serve more from cache, including static and event-driven content. This improves your cache hit ratio, resulting in better user experiences.


Network map

North America

  • Ashburn 3X
  • Atlanta 3X
  • Boston
  • Chicago 3X
  • Columbus
  • Dallas 2X
  • Denver
  • Houston
  • Jacksonville
  • Los Angeles 2X
  • Miami
  • Minneapolis
  • Montreal
  • New York 3X
  • San Jose 2X
  • Seattle
  • Toronto
  • Vancouver

South America

  • Bogota
  • Buenos Aires
  • Curitiba
  • São Paulo
  • Santiago
  • Rio de Janeiro

Europe

  • Amsterdam
  • Copenhagen
  • Dublin
  • Frankfurt 2X
  • Helsinki
  • London 3X
  • Madrid
  • Manchester
  • Milan
  • Oslo
  • Paris
  • Stockholm
  • Vienna

Africa

  • Cape Town
  • Johannesburg

Asia

  • Chennai
  • Dubai
  • Hong Kong
  • Mumbai
  • New Delhi
  • Osaka
  • Singapore
  • Tokyo 2X

Australia & New Zealand

  • Auckland
  • Brisbane
  • Melbourne
  • Perth
  • Sydney
  • Wellington

Coming soon

  • Bangkok, TH
  • Berlin, Germany
  • Portland, OR
  • St Louis, MO
  • Tel Aviv, Israel
  • Warsaw, PL
  • Washington, DC

Network map last updated: December 31, 2019

Global capacity

74 Tbps of connected global capacity

Capacity last updated December 31, 2019

Fastly cache server specs

  • 24 TB of SSDs
  • 768 GB of RAM
  • 4 x 25 Gigabit Ethernet

Performance by the numbers

  • 150ms mean purge time to clear cached content globally, as of December 31, 2019
  • ~ 5 sec to change a config through the web API
  • Sub-millisecond TTFB on the 99th%

Peering with Fastly

Fastly peers with other Internet Service Providers (ISPs) and Content Networks with IPv4 and IPv6 connectivity on Autonomous System (AS) 54113 for the purpose of exchanging traffic between these networks. To learn more about peering with Fastly, visit our peering page.

The power of the edge

Our real-time “Tachometer” displays Requests per second, Cache hit ratio, Hit time, and Bandwidth based on aggregated requests and bandwidth across our network. These figures include usage and services from all of our customers, as well as trial, developer, nonprofit and open source program, and other non-paying accounts. These figures also include usage from our own internal health check requests. As such, these figures are only for informational purposes to show the scale of our network and do not reflect or predict our revenue or financial performance. You should not rely on our Tachometer to evaluate our business or expected financial results.

Requests per second

This is a count of every request across our network in the last second.

Cache hit ratio

This is the ratio of cacheable content served from cache to cacheable content that needed to be fetched from a customer’s origin. Cacheable content is defined as content that has not been explicitly marked to always be fetched from the origin.

Hit time

The average amount of time it takes to serve a piece of content that is cached on one of our servers.

Bandwidth

This is the current aggregate size of all content served from all our servers in the last second.

  • Fastly POP
  • Multiple POPs
  • Planned POPs