S3

Question 1

What S3 stands for?

Answer 1

Simple Storage Service

Question 2

What is S3 bucket?

Answer 2

It is a folder/container of objects/files

Question 3

Why the S3 bucket needs to be globally unique?

Answer 3

Because a CNAME will be create such as https://.s3.amazonaws.com and https://.us-east-1.amazonaws.com

Question 4

What is the structure of the object within S3 bucket?

Answer 4

Key, Value, Version ID, Metadata and sub resources (Access Control Lists and Torrent)

Question 5

How S3 keeps data consistency?

Answer 5

If you write, you can read immediately. But if you delete/update, you may get the older version.

Question 6

What is the availability of S3?

Answer 6

99.99% for the S3 platform. 99.9% for Amazon, and 99.999999999 durability for S3 information

Question 7

What are the different S3 storage classes?

Answer 7

S3 standard: Data is stored in multiple device facilities
S3 Infrequently Accessed: Lower cost than S3 standard, but charges on retrieval fee
S3 One Zone - Infrequently Accesses: Cost effective one zone
S3 - Intelligent Tiering: Optimize costs by automatically moving data to most cost-effective (IA) access tier without performance impact
S3 Glacier: Cheaper than on-premise. Retrieval times from minutes to hours;
S3 Glacier Deep Archive: Lowest cost class where a retrieval time of 12h is acceptable

Question 8

How S3 is charged?

Answer 8

Storage (volume)
Request (traffic)
Storage Management (classes)
Data Transfer
Transfer acceleration
Replication

Question 9

What is the maximum file size?

Answer 9

5TB, but for a single PUT it is 5GB. In case the file size is bigger than 100MB, multipart upload is suggested

Question 10

What is the limite for S3 storage?

Answer 10

Virtually unlimited

Question 11

What is Transfer acceleration?

Answer 11

It is when the data is replicated using the AWS backbone (fastest network)

Question 12

How to protect object deletion?

Answer 12

It is possible to enable MFA for Deletion

Question 13

What is Access Control List?

Answer 13

Permissions on bucket or object level

Question 14

Can I have a bucket that has different objects in different storage classes?

Answer 14

Yes, you can have a bucket that has different objects stored in S3 Standard, S3 Intelligent-Tiering, S3 Standard-IA, and S3 One Zone-IA.

Question 15

How to control access control?

Answer 15

Use Bucket Policies for buckets;

Use Access Control Lists for objects;

Question 16

How versioning works?

Answer 16

All writes and deletes of an objects are versioned. It is great as backup tool. Once enabled, it can't be disabled, only suspended (disabled for new objects only);
The visibility (make it public) needs to be done per object version;

Question 17

How lifecycle management can be used?

Answer 17

Transition objects/versions from one tier to another based on number of days

Question 18

What kind of governance S3 has?

Answer 18

Using S3 Object lock, the bucket/object is written, but can’t be deleted. It comes in two modes: governance-users can’t delete or alter unless they have special permissions; compliance mode: not even root can change the object;

Question 19

How to increase performance when using S3?

Answer 19

Spread objects among multiple prefixes/folders. In general, it accepts 5500 PUT/second.
Use multipart upload: Recommended for 100MB+ files; Mandatory for 5GB+;
S3 byte-range fetches for downloading faster or just a part of the file

Question 20

What is S3 or Glacier Select?

Answer 20

It allows queries to be done against compressed CSV zip files in S3 with a performance improvement of 400%.
This also allows queries to be executed on data that is stored in Glacier (low cost data analysis)

Question 21

Are the existing S3 bucket objects replicated when a replication rule is just created?

Answer 21

No. Only new objects versions will be replicated. Interesting to remember that the object can be replicated to a different account with a different storage class.
Replicated object permission changes are not replicated

Question 22

What is the pre-requisite for enabling S3 replication?

Answer 22

Versioning

Question 23

What is Snowball?

Answer 23

petabyte-scal data transport solution with 1/5 cost of high-speed internet. It can import/export from/to S3

Question 24

Snowball vs Snowball edge?

Answer 24

Snowball is a storage with 50TB or 80TB. Snowball edge comes with 100TB and it also provides computing capability. It can be clustered to run workloads even on places without internet access.

Question 25

What is the avilability for each S3 class?

Answer 25

The S3 Standard storage class is designed for 99.99% availability, the S3 Standard-IA storage class is designed for 99.9% availability, the S3 One Zone-IA storage class is designed for 99.5% availability, and the S3 Glacier and S3 Glacier Deep Archive class are designed for 99.99% availability and SLA of 99.9%.