Fly.io compared to Amazon Lambda @ Edge

Fly.io
Versus
Amazon Lambda @ Edge

Features

Edge Features of Fly.io compared to Amazon Lambda @ Edge
Fly.ioFeaturesAmazon Lambda @ Edge
Functions / Serverless
Functions supported languagesjavascript, go, C/++, .NET, Node.js, PHP, python, ruby
Worker.js Environment
Docker supportYes, through EC2 Container Registry (ECR)
Docker private registry
Kubernetes support
Managed Kubernetes
Availability regionsAll compute pops
Default Memory (MB)128
Maximum Memory (MB)128
No limitExecution Time (ms)5,000
No limitMaximum Execution Time (ms)30,000
No limitRequest Payload (MB)50
No limitResponse Payload (MB).04
Unsupported Paid Feature Supported Unknown

Descriptions


Fly.io


Fly is an opensource runtime for edge apps, Fly offers a paid hosted version of their product as well. This provider will focus on the paid hosted version. Fly.io is one of the few companies and products that has a full range of tools alongside their actual product, which makes them an ideal candidate for curious developers and new companies to experiment.

Everything they do is docker based, and their networking abstraction is top notch, it’s clear Fly will play a major role in turning the space into a more user-friendly one. Fly’s ambitious mission to make application distribution as ubiquitous as CDNs sets the bar for any contenter willing to compete.


Amazon Lambda @ Edge


Amazon Lambda at the Edge functions introduced serverless cloud computing to the masses as early as 2014.

Being the first with a massive user-base has set up Amazon for great success, it took a few years for competitors to offer similar functionality and to actually call the FaaS space a new chapter in cloud compute in general.

Amazon’s Lambda@Edge is Amazon’s first Edge Compute product, however Amazon recently released CloudFront Functions, which brings the compute a lot closer to the end-user.

With Amazon’s gigantic scale, a shift has started with companies that are on the AWS platform to move more and more of their infrastructure to serverless.