It started as a meme, but it turned into a real post on “The 17 Ways to Run Containers on AWS.” Apparently my list continues to be a source of amusement inside of AWS.
Given that I do prefer to give the people what they want, I’d like to talk about 17 more ways to run containers on AWS.
“Now hang on a second, Corey,” you might be thinking, “re:Invent is in a couple of months. If you wait until then, AWS is all but certain to release at least a couple more services that can be pressed into container-running duty.”
I appreciate how thoughtful you are, but we don’t actually need to wait for re:Invent. So if it’s all the same to you, I’ll list the entire second set of 17 services right now.
1. AWS Batch
A bunch of folks wrote in to tell me that I forgot about AWS Batch in the first article. This is patently untrue! I remembered Batch, but it’s SUPER hard to run containers with it — despite that being a stated goal of the service.
2. AWS Nitro Enclaves
Nitro Enclaves are isolated compute environments that protect and securely process highly sensitive data. The way you build enclave image files is, of course, via Docker images.
3. AWS Step Functions
While people generally consider Step Functions to be a hilariously over-complicated way of invoking Lambda functions, they also absolutely support container tasks as well.
4. VMware Cloud on AWS
You can use VMware Cloud on AWS to run Kubernetes via whatever the hell a “Tanzu” might be, and in fact, it’s something actively recommended by VMware. This apparently solves the problem with Kubernetes. What problem might that be? According to VMware, it’s “We don’t make money when you run Kubernetes yourself.”
5. Amazon App Stream 2.0
This little-known service lets you stream desktop applications. This, of course, includes Docker Desktop so of course it counts.
6. Amazon Nimble Studio
Nimble Studio is a managed video studio solution. It automatically provisions and maintains a studio workstation inside of an EC2 instance, inside of which you can run containers.
7. AWS Marketplace for Containers
Not only can you use this service to run containers, you can buy them as well.
8. Amazon SageMaker
Loath though I am to suggest using this Machine Learning® monstrosity for anything, you don’t have to twist that hard to get this thing to run containers. It is, in fact, AWS’ recommended approach.
9. AWS Outposts
“Servers in a rack that get shipped to your facility” open up many different possibilities for customers, but, really, I just want to use one for a second to run a container.
10. Amazon WorkSpaces
Amazon WorkSpaces are virtual desktops running inside of AWS. “Running containers” is a standard part of many folks’ usual workflows with these things.
11. AWS CloudShell
AWS recently announced CloudShell, which, unlike everything else on this list, is completely free. Sure, AWS manages the container, but it does, indeed, run the container and let you do an awful lot inside of it.
12. AWS Device Farm
Device Farm is one of those “are you messing with me or not?” services, based upon nothing other than the description. “A bunch of Android and iOS devices sitting in a rack somewhere for testing mobile apps? Really?” Well, it takes some doing, but you can indeed get Docker running on Android devices.
13. Amazon GameLift
GameLift apparently runs gaming servers for you. It natively supports containers as one specific architectural choice you can make, and it suggests leveraging Spot Instances for it as well.
14. AWS Cloud9
Cloud9 (“like GitHub CodeSpaces, only terrible!”) runs its own IDE on top of either a provisioned EC2 instance or your own environment. You can absolutely leverage containers for this; many folks do intentionally do this as part of their development process.
15. and 16. AWS Snowball Edge and AWS Snowcone
The Snowball Edge (“please do not look either of those words up on Urban Dictionary”) and the Snowcone (“the worst Kindle e-reader that Amazon has ever made”) both support running EC2 instances on them — and, naturally, you can then run containers within them.
17. Amazon RoboMaker
As of last week, RoboMaker now explicitly offers support for container images. Aren’t you just so very glad that they made that change?
Please don’t make me write another containers post
So there you have it: 17 more ways to run containers on AWS. I really hope there isn’t going to be a third post in this series, but AWS is going to AWS.
If you simply can’t contain yourself, feel free to scream at me on Twitter about this at your convenience.
