This article describes the features and functions of the Fastly CDN.
Fastly CDN works quickly by storing website content on servers around the world and delivering that content to users quickly.
Fastly uses 'Varnish', an open-source web application accelerator, which makes the transmission of its content (content objects) more efficient by automatically storing a copy in a temporary intermediate location.
It tracks the geographical location of each user and ensures that the user is connected to the nearest server. This speeds up the site by reducing the time spent waiting for data to be sent from the server to the user.
Fastly also have full control over how and when content from the server is stored.
You can set a time to live (TTL) for any path on your site and use the PurgeAPI to instantly disable or purge any path on your site.
By using these tools, a site page only needs to be generated once for millions of page views. This reduces running costs for user requests and server usage fees.
Fastly CDN offers a number of advanced features to help monitor how data is accessed and customize content delivery.
Better control over when and how content is updated.
Instead of waiting up to 24 hours for data to change at the edge, you can update data when and how often you need to.
Provides a top-level view of network and site performance.
Compiles relevant data on all traffic every second into easy-to-read reports.
It changes the way requests are routed, the headers sent, and the content cached.
If there is a problem with the server, you will know about it immediately.
It can be quickly and easily configured to send information from the server to anywhere in the required format.
Almost all aspects of HTTP requests and responses can be modified. A specially configured VCL file can be uploaded to your account.
An API exposed by Fastly allows you to handle program configuration, handle basic configuration tasks and create scripts to create your own management views.
How It Works
Fastly provides CDN services from key access points to the Internet called 'points of presence' (POPs). Then connections to the Internet allow POPs to be quickly placed in locations where network transit times are reduced when delivering content to end-users.
Each POP has a cluster of Fastly cache servers that, rapidly deliver content objects from the cache location closest to each end-user when they request a content object.
Fastly's cache only receives and processes end-user requests for content objects such as:
■ Which objects to cache
■ How long and who can access the objects
■ Whether to encrypt objects when transmitting over the Internet
■ When to remove objects from the cache service
Decisions are made by specifically configuring Fastly's CDN services using requirements such as above. This configuration process is called 'provisioning'.
To provision Fastly's CDN services, you need to identify the application servers that serve the original content objects for each of the various domains.
The application server can be a physical server in a data center or hosting facility, an application running on a cloud service such as AWS or GCP, or any combination of these.
Fast reference is made to these source servers in the same sense as 'origin servers' and 'back-end servers'.
In addition to using its own servers as a source, Fastly CDN also supports the use of various cloud storage services as an origin, such as Amazon S3 and Google Cloud Storage (GCS).
We hope that this article will be of interest to Fastly's CDN.
This blog post is translated from a blog post written by Yuya Ohara on our Japanese website Beyond Co..