What the cloud sells: services and regions

Cloud computing means renting ready-made services, such as file storage, from a provider over the web. Most apps you use store their data this way, so knowing how a call works is basic to the job.

A shop lets customers upload a photo of a faulty item. The photo is stored in an S3 bucket, a place for files in the cloud. Your code does not log in to a machine to save it. It sends a request over the web and reads an answer.

A cloud provider runs big computers and rents out what they do. Each rented tool is a service. S3 stores files. SQS holds messages in a queue. DynamoDB stores items in tables.

You use a service by making an API call: a request that names the call, such as list_objects_v2, and the details it needs. The answer comes back as data.

call = {"service": "s3", "operation": "list_objects_v2", "Bucket": "shop-photos"}
answer = {
    "Name": "shop-photos",
    "KeyCount": 2,
    "IsTruncated": False,
    "Contents": [
        {"Key": "faulty/kettle.png", "Size": 2048, "StorageClass": "STANDARD"},
        {"Key": "faulty/toaster.png", "Size": 3072, "StorageClass": "STANDARD"},
    ],
}
print("service:", call["service"])
print("call:", call["operation"])
print("keys:", answer["KeyCount"])
for item in answer["Contents"]:
    print(item["Key"], item["Size"])
# Change the Size of the toaster to 5000, then press Run again.

Not every call works. A key may be missing or a bucket may not exist. Then the service answers with an error instead: a Code that names the problem and an HTTP status.

Your code must look for the error first. Code such as "Contents" in answer only works when the call worked.

good = {"KeyCount": 1, "Contents": [{"Key": "readme.txt", "Size": 5}]}
bad = {"error": {"Code": "NoSuchBucket", "HTTPStatusCode": 404}}
for answer in [good, bad]:
    if "error" in answer:
        print("failed:", answer["error"]["Code"], answer["error"]["HTTPStatusCode"])
    else:
        print("worked, keys:", answer["KeyCount"])
# Change "NoSuchBucket" to "NoSuchKey", then press Run again.

A provider has data centres in many places. A region is one area of them, such as eu-west-2, which is London. You choose the region when you make a bucket, and the files stay there.

Pick a region near your users, so answers arrive quickly. The answer that lists your buckets also shows the region of each one.

answer = {
    "Buckets": [
        {"Name": "shop-photos", "BucketRegion": "eu-west-2", "BucketArn": "arn:aws:s3:::shop-photos"},
        {"Name": "shop-backups", "BucketRegion": "us-east-1", "BucketArn": "arn:aws:s3:::shop-backups"},
    ],
    "Owner": {"DisplayName": "shop", "ID": "owner-id"},
}
for bucket in answer["Buckets"]:
    print(bucket["Name"], "is in", bucket["BucketRegion"])
# Print only the buckets in eu-west-2: add an if inside the loop.

The provider looks after the buildings, the machines and the service software. You look after what you put in it and who may use it.

This is shared responsibility. If you let anyone read a bucket, that was your choice. If a data centre loses power, that is the provider's job to handle.