Over the past three months, we’ve increased network capacity by 40% through a variety of transit augments and new public and private interconnection arrangements. We are now present at Internet Exchange Points (IXPs) in Amsterdam (AMS-IX), Boston (BOSIX), and Toronto (TORIX) in addition to existing locations across the North America and Asia Equinix Exchanges, LONAP in London, BBIX in Japan, Megaport and WAIX throughout Australia, and IX.br in São Paulo.
AMS-IX is notable as it’s the location of our first 100GbE ports on the Fastly network, which allow us to maximize our hardware resources (as well as being really cool from a technology standpoint).
Our network team attended NANOG in Chicago, sporting the iconic red shirts with our ASN54113 number on the back. You can learn more about Fastly's Peering Policy at https://www.fastly.com/peering and our peering footprint at PeeringDB: http://as54113.peeringdb.com/.
This quarter we focused on augments to our cache server footprints across the United States. Atlanta, Dallas, Chicago, New York, Virginia, and Los Angeles were all given upgrades with new hardware (and more of it!) to increase our caching capacity at some of our most heavily trafficked locations.
At the urging of our Canadian-born network and data center team members, we expanded into Canada with the launch of our Why-Why-Zed (YYZ) POP in Toronto, which leads to lower latency and improved performance for Canadian and US traffic near the border. Our infrastructure team employees living in Massachusetts and New Hampshire also stepped up to the regional challenge by launching a new POP in Boston.
In Europe, we doubled the cache server capacity at one of our London data centers and installed cache drive upgrades in our Amsterdam servers. More servers and more drives mean higher caching rates across our European footprint.
And finally, as we announced on the status page, we have scheduled our new POP in Dubai, UAE (POP code FJR), to go live in July. This will give us a strategic site to improve Eastern European and Middle East performance.
We introduced a new Drupal 8 module, which utilizes Drupal’s new cache tags and allows us to invalidate cache based on specific cache tags. Our Fastly 8.x module utilizes Drupal core’s CacheTagsInvalidatorInterface to purge specific cache tags allowing for all content associated with the cache tag to be invalidated. We have also applied additional Bugfixes and community patches to Drupal 7x.
We recently released several major updates to our Heroku add-on integration. Heroku customers now have full availability of the Fastly UI, which can be accessed within their Heroku dashboard via single-sign on (SSO). We also launched Heroku-Fastly, a new plugin for Heroku’s powerful command-line interface (CLI). The initial release of our CLI plugin for the Heroku toolbelt includes two commands: purge and TLS provisioning of custom domains registered with Fastly.
To better serve developers who host lower-traffic sites, we now offer a $25/month pricing plan that includes 100GB of bandwidth and 1M requests — our lowest price to date. Additionally, Heroku pricing plans at the “Faster” level or above now include one or more TLS domain.
A Fastly extension for Magento 2 is now available on our GitHub page and the Magento Marketplace. Once again we partnered with Phoenix Media, a systems integrator with deep Varnish and Magento expertise, to rewrite our Magento 1 extension for the Magneto 2 platform so customers can take advantage of Fastly-specific features such as surrogate keys, GeoIP, crypto extensions, and Origin Shield.
Audit Logs are now accessible via the API, and will help you keep track of changes to your service, including which changes were made, when, and by whom. Audit Logs will also help you secure configurations, improve monitoring, and conduct better troubleshooting and debugging when abnormal behavior arises. Audit Logs are API-only for now. For a full list of supported Audit Log events, check out the API documentation.
User-level API tokens are now in Limited Availability and can be accessed via our API. Tokens are unique authentication identifiers for users and applications that are authorized to interact with your service. Since tokens can be scoped to a single service and restrict access to purge functions, they give you more control and visibility of who accesses and uses the Fastly API. And a single user can have multiple tokens, which means you can rotate or revoke tokens without taking services offline or updating other API integrations. For more information about user-level API tokens, check out the documentation.
Our Edge Access Control Lists (ACLs) are in General Availability. Our newest addition to our edge family, they provide a way to programatically create, update, and delete ACLs without incrementing your Fastly configuration version. For more information, check our documentation.
IPv6 support is in Limited Availability. IPv6 was developed by IETF to deal with the long-anticipated problem of IPv4 address exhaustion, and is intended to eventually replace IPv4. IPv6 also offers many enhancements to IPv4, such as better performance in mobile networks. If you are interested in enrolling in the LA program, email email@example.com to get it turned on for your account.
Fastly two-factor authentication (2FA) is in General Availability. You are now able to turn this feature on/off for your entire company. We’ve made it possible for you to enforce these settings at the company and individual level. Individuals users will not be able to log into Fastly without setting this up, greatly reducing support and monitoring time. For more details, check out our documentation.
We are excited to announce that our HTTP/2 Limited Availability (LA) program is here! Fastly implementation of the protocol will support all standard features, as well as options like server push, including:
For more information on how to sign up, check out our blog on HTTP/2 updates.
In February of this year, we announced our revised deprecation plan for TLS 1.0 and 1.1. As part of that plan, we promised to provide customers who wish to enforce stricter security requirements an opportunity to migrate to hosts that only support TLS connections via the TLS 1.2 protocol ahead of our planned deprecation schedule for the rest of the network. We’re happy to announce that you can now request migration to these TLS 1.2-only hosts if you’ve purchased a paid TLS option. For more details check out our latest blog.
We’ve released OTFP Video revalidation functionality, which returns 304s rather than repackaging on time to live (TTL) expiration, also called ETag validation. This reduces origin egress and midgress traffic and lowers Time To First Byte (TTFB) on expired objects, improving the viewing experience.
Fastly is Wowza’s newest Tech Alliance Partner for delivery. Combining Wowza’s robust transcoding capabilities with Fastly’s secure delivery helps joint customers simplify their live streaming workflows while delivering exceptional viewing experiences. To learn more about this partnership, check out Wowza’s blog post.
We have a new dedicated space for online video streaming guides and tutorials for our OTFP product including descriptions, available features, and other helpful information.
This past quarter, Fastly had a presence at the following events:
Our engineers also spoke at CraftConf, QCon, Software Architecture Conference, OSCON, ACSC, RenderConf, Interop, SRECon, Global Peering Forum, and more.
Check out our events page to keep track of conferences we’ve been to and ones we’ll be attending in the coming months.
We’re continuing our efforts to support open source projects by donating our services. Here are some projects that have recently started using Fastly:
As always, if you have an open source project that can use Fastly services, please reach out to firstname.lastname@example.org. Want to chat with Fastly engineers and other customers using our product? Check out our Community Forum.