Chapter 8 - Databases

Question 1

RDS databases (Amazon Relational Database Solution)

Answer 1

• RDS database types come in SQL server, Oracle, MySQL,PostgreSQL, MariaDB, and Amazon Aurora.
• RDS is for online transaction processing workloads.
• So, this is where you are basically processing lots of small transactions like customer orders, banking transactions, payments, and booking systems.

Question 2

Redshift

Answer 2

• For data warehousing
• OLAP (online analytical processing) tasks like analyzing large amounts of data reporting and sales forecasting.

Question 3

Read Replicas

Answer 3

• This is where you’re scaling your read performance.
• It’s primarily used for scaling
• Not for disaster recovery.
• It requires automatic backups to be turned on.
• So, if you can’t for some reason create a Read Replica of your primary database,it just means you haven’t turned on automatic backups.
• Multiple Read Replicas are supported.
• So, you can have up to 5 Read Replicas per MySQL, MariaDB, PostgreSQL, Oracle, and SQL server.
• Read Replicas read-only copy of your primary database in the same availability zone,
• cross availability zones or across regions is available.
• And of course you use this to increase or scale your read performance.
• It’s great for read heavy workloads and it takes the load off your primary database for read-only workloads.

Question 4

Aurora

Answer 4

• Amazon’s proprietary database that’s something that they have created themselves
• It’s compatible with MySQL, as well as PostgreSQL.
• It always has 2 copies of your data in each availability zone with a minimum of 3 availability zones.
• So, you always have 6 copies of your data.
• So it’s very, very redundant.
• You can also share Aurora snapshots with other AWS accounts

You have 3 types of replicas available with Aurora.

1. You’ve got a Aurora replicas themselves,
2. You can also create MySQL replicas,
3. And PostgreSQL replicas.

• An automated failover is actually only available when you have Aurora replicas.
• Aurora has automated backups turned on by default.
• You can also take snapshots with Auroras
• You can share these snapshots with other AWS accounts.

Question 5

DynamoDB

Answer 5

• it’s stored on SSD storage
• it’s spread across 3 geographically distinct data centers
• You can have eventually consistent reads which is what you get by default,
• You can also have strongly consistent reads.

Question 6

DynamoDB transactions

Answer 6

• This is where you have multiple all-or-nothing operations.
• So, it’s good for things like financial transactions or fulfilling orders.

You have 3 options for your reads

1. Eventual consistency,
2. Strong consistency and then
3. Transactional consistency

You have 2 options for your writes

1. Standard
2. transactional
3. You can have 25 items using DynamoDB transactions or 4 MB of data.

ACID requirements
DynamoDB transactions provides developers

1. atomicity
2. consistency
3. isolation,
4. durability across one or more tables within a AWS account or AWS region.

If you see something about ACID requirements with DynamoDB, you need to use DynamoDB transactions.
This is basically means, you’ve got all-or-nothing transactions.

Question 7

DynamoDB On-Demand backups and restore

Answer 7

• This is where you can backup your DynamoDB database with full backups at any time.
• It has zero impact on your table performance or availability.
• And you have consistent within seconds
• And your backups are retained until deleted.
• And it operates within the same region as the source tables.
• When you’re creating a DynamoDB backup, it will be in the same region as the source table.

Question 8

DynamoDB Point-in-Time Recovery

Answer 8

• It protects against accidental writes or deletes
• You can restore your DynamoDB database to any point in the last 35 days
• The backups are incremental.
• It’s not enabled by default.
• However, you will have to go and turn it on.
• And the last restorable rate is going to be 5 minutes in the past.

Question 9

DynamoDB streams

Answer 9

this is where you can maintain first in first out records of your data.

So as you go and add transactions -

• basically it’s given a sequence number.
• These sequences are stored in a stream,
• and the data’s broken up into shards.
• Every shard is stored for 24 hours
• and your stream records are going to consist of things like inserts, updates, and deletes.
• And you can combine this with a Lambda function to add functionality like stored procedures.

So, it’s just time audit sequences of item level changes in a table.

Question 10

global tables

Answer 10

• this is where you have managed Multi-Master, Multi-Region Replication.
• It is designed where you’ve got applications that are globally distributed.
• You need to have DynamoDB streams turned on.
• So, it’s based off DynamoDB streams.
• Basically allows you to have Multi-region redundancy for disaster recovery or high availability.
• And you don’t have to rewrite their application.
• This is all built in functionality within DynamoDB.
• And your application latency typically will be under one second.

And the key thing here is to remember

• if they’re talking about if you want to add redundancy to DynamoDB,you need to turn on global tables.
• If global tables isn’t turning on, you need to make sure that you’ve enabled DynamoDB streams.