Q3 2019

Hi Fastly customers + friends,

In Q3 2019, we moved Media Shield and Segmented Caching to GA, improved the UI of the “Add a domain” feature within the Fastly app control panel, restructured the navigation of Fastly’s documentation site based on customer feedback and usability testing, added support for the latest version of Terraform, and added Elasticsearch as an endpoint to our Real-Time Log Streaming feature. Continue reading for details on this quarters’ releases.

Table of Contents

• Network updates


• Media Shield GA


• Segmented Caching GA


• Site Configuration and Onboarding


• Completed Terraform work


• Elasticsearch logging endpoint LA


• Varnish & VCL updates

Network updates: 66 global POPs, 58 Tbps of edge capacity

We now operate 66 global POPs. We recently deployed sites in Bogota, Jacksonville, Singapore, Minneapolis, and Manchester. Our new sites and upgraded POPs extend our total connected edge capacity to 58 Tbps as of 9/30/19.

Media Shield GA

Media Shield went GA on September 13, 2019.

Media Shield helps customers reduce the costs of live streaming video in multi-CDN architectures by consolidating multiple requests for the same content from different CDNs down to one request to origin. This decreases the number of requests your infrastructure needs to process by several orders of magnitude, and helps save on network egress costs. Media Shield has been used to empower multi-CDN strategies for the largest live sports events in the world, facilitate live linear/vMVPD services and partner with online video platforms (OVPs).

Observability, the ability to know what is happening inside the tech stack, is critical to the success of a live event. Media Shield allows customers to have real-time observability by using real-time log streaming to the endpoints of your choice. This allows you to detect and remediate issues with origin, Media Shield or edge infrastructure as soon as they occur. And we’ve responded to customer needs by adding new endpoints over with Honeycomb, Splunk, Azure Blob Storage, and Kafka in general availability and Rackspace CloudFiles, OpenStack and Log Shuttle during limited availability. This functionality is core to our platform and key to empowering customer video delivery operations teams. While lots of customers take advantage of real-time logging, many want an easy-to-use API endpoints with stats that might trigger a deeper dive into logs. To do this, customers leverage the Fastly Real-Time Analytics API. We continue to invest in new features to make live video delivery faster, easier, and more scalable.

Segmented Caching GA

Segmented Caching went GA on September 24, 2019.

Segmented Caching allows you to cache files of any size and improves the efficiency of byte-range requests. Segmented Caching works by breaking files into smaller, 1MB segments in Fastly’s cache, then recombining or splitting these objects to respond to arbitrary size byte-range requests from clients.

Now customers can serve up gaming content, videos or software downloads in smaller file sizes, and we can fetch just what the user wants in terms of content. For example, if the user wanted to watch the beginning of a video and not the entire video, we can stop serving content once the user stops requesting it. This can help save on egress costs since we’re serving only the content that the user requested, rather than the whole file. And it’s efficient for Fastly as well, because we are dividing these files into smaller sizes. Any file size can be supported, making it easily scalable. This feature is available to all customers leveraging Fastly for streaming and content delivery.

Site Configuration and Onboarding

As part of our ongoing effort to simplify the onboarding experience, we updated the “Add a domain” feature in the Control Panel to be an inline form vs. a lightbox modal. This reduces friction to the process of adding a domain and makes getting up and running on Fastly that much easier.

Docs Site IA Revamp

We are pleased to introduce a new taxonomy for the docs.fastly.com site. Taking into account customer feedback via usability testing and surveys, the new information architecture improves discoverability and orientation within the Docs site. The new navigational structure provides better entry points for customers who are new to Fastly and provides clear next steps in the learning journey. Visit docs.fastly.com today to get a feel for the new site experience.

Completed Terraform Provider Update

Terraform is a popular developer tool for building, changing, and versioning infrastructure safely and efficiently. With Terraform, joint customers can treat their Fastly configurations as code and automate the management of their infrastructure more than ever before.

Fastly now offers support for the latest version of Terraform (0.12.0) that, among other things, allows customers to iterate over configuration. We are also proud to provide support for dynamic snippets, versionless ACLs and edge dictionaries, which has been a common request. We also added support for Microsoft Azure Blob Storage, Splunk logging and version and service comments.

Elasticsearch Logging Endpoint LA

Real-Time Log Streaming to Elasticsearch went LA on September 3, 2019.

Fastly's Real-Time Log Streaming feature can now send log files to Elasticsearch. Elasticsearch is a distributed, RESTful search and analytics engine capable of storing, searching, and analyzing large volumes of data in near real-time. Many companies are using Elasticsearch to power their applications that have sophisticated search features and requirements. This logging endpoint is currently in Limited Availability and disabled by default. To enable this endpoint for your account, contact support@fastly.com.

Varnish & VCL Updates

The following varnish and/or VCL updates were deployed to the worldwide Fastly edge cloud fleet in Q3 2019:

• The new std.strrev() function reverses a given string.


• 16 new VCL variables were made available in support of the Segmented Caching feature.

These updates relate to string manipulation and controlling range requests via Segmented Caching.