Serverless

Click to play video

"Serverless" is an architecture (and let's be honest, a buzzword) that refers to a system where you don't have to manage the servers on your own.

Serverless is largely misunderstood due to the dubious naming. It does not mean there are no servers, it just means they're someone else's problem.

You'll often see "Serverless" used to describe services like AWS Lambda, Google Cloud Functions, and Azure Functions. And that's true, but it refers to "serverless" in its most "pure" form: serverless compute.

AWS S3 was actually one of the first "serverless" services, and is arguably still the most popular. It's not serverless compute, it's serverless storage. You don't have to manage/scale/secure the servers that store your files, AWS does that for you.

Instead of going to a local file system, your server makes network requests to the S3 API to read and write files.

Assignment

Make a S3 bucket through the AWS console called tubely-<random_number>. Replace <random_number> with a "random" (just choose one) number to the end of the bucket name to ensure it's unique. For example, tubely-56841, but choose your own number.

The bucket name has to be tubely-<random_number> to pass the tests.

Uncheck "Block all public access" when creating the bucket.
Leave bucket versioning off.
Leave default encryption on with managed keys.
Leave object lock disabled.

Use the AWS CLI to ensure the bucket is there.

aws s3 ls

Now that we have a bucket, we still need to configure its permissions to control who can access the files and how, typically through a bucket policy or object-level ACLs.

Go to your bucket and configure the bucket policy (under the "Permissions" tab). Copy-paste the following, replacing BUCKET_NAME with your bucket.

{
  "Version": "2012-10-17",
  "Statement": [
    {
      "Effect": "Allow",
      "Principal": "*",
      "Action": "s3:GetObject",
      "Resource": "arn:aws:s3:::BUCKET_NAME/*"
    }
  ]
}

This will give read-only access to anyone with the precise URL of an object. Notably, this doesn't allow listing the objects in the bucket, only reading them if you know the exact URL.

Run and submit the CLI tests.