Databases Flashcards
Relational Database Types
SQL Oracle MySQL PostgreSQL Aurora MariaDB
Non-relational databases on AWS
DynamoDB
DynamoDB Database Parts
Collection = table
Document = Row
KV pairs = Field/data
OLTP SF
Online Transaction Processing
OLAP SF
Online Analytics Processing
OLTP is
Pulling up data, a basic request
OLAP
Analysis of a large volume of information on a database, it can use a lot of resources. Usually you̥ll use a copy
Elasticache is
A deployed in-memory cache in the cloud. It caches the most consistently queries fields to make the database faster.
DMS SF
Database Migration Services
DMS
Allows migrations to databases and does transformation, compression and parallel transfers and converts to formats of target database
RDS OLTP Types
SQL
MySQL
PostgreSQL
Oracle
Aurora
MariaDB
OLAP Types
Redshift
NoSQL DB types
DynamoDB
Elasticache Types
Memcached
Redis
DynamoDB is
Fast flexible NoSQL database for document and key-value data models
Why use DynamoDB?
SSD Storage
Spread across 3 geographically distinct data centres
Dynamo Eventual Consistent Reads
Default setting. Consistently copies all data within a second to other data centers. Give best read performance
Basic inclusions of DynamoDB
Tables
Items (rows)
Attributes (columns)
Levels of nesting supported in Dynamo
35
Dynamo pricing for Write Throughput
$0.0065 per hour for every 10 units
Dynamo pricing for read throughput
$0.00065 per hour for every 50 units
Dynamo storage cost
first 25 GB stored per month is free
$.25 per GB per month there after
Dynamo Strongly Consistent Reads
Reflects all writes that receives a successful response prior to the read
DynamoDB the 2 types of primary keys
Single attribute
Composite
DynamoDb Unique ID
(think unique ID)
aka partition key
DynamoDB Composite Key
Think unique ID and a date range
Partition Key & Sort key composed of 2 attributes
DDB Partition Key
Dynamo uses the partition keys value as input to an internal hash function.
Must be unique!
DDB Two items with the same partition key
Must have a different sort key
All items of the same partition key are stored together sorted by sort key value
DDB Local Secondary Index
SAME partition key, different sort key
Can only be created with creating a table
DDB Global Secondary Index
Has different partition and sort key
Can be added at any time
DDB Streams
Captures modifications
- Any new items
- Tracks changed items
- Holds deleted items
Stores for 24 hours
Useful for sending an on boarding email
DDB Query
Query using only primary key value. Can also use sort key (usually a date)
Types of indexes
Local secondary
Global secondary
Scan vs Query
Query only finds by primary key value (can use sort)
Scan looks at all the data and filters out (slower!)
ProjectionExpression
Parameter for scanning so that the query only return some of the attributes, not all
To get query results reverse sorted
Set ScanIndexForward to false
To limit what scan shows
Use a projection expression
Units of read provisioned throughput
Reads are rounded up to 4kb increments
Consistent reads consist of 2 reads per second
Strongly consistent reads are 1 read per second
Unit of write provisioned throughput
All writes are 1 Kb
All write consist of 1 Kb per second
read throughput calculation
(Size of read rounded to nearest 4kb chunk/4) * no of items
eventually consistent read throughput calculation
(Size of read rounded to nearest 4kb chunk/4) * no of items
Then divide by 2
You have an application that requires to read 10 items of 1 kb per second using eventual consistency. What should you set the read throughput to?
(1kb4/4) = 1 read unit per item
110 = 10
10/2 = 5 units of read throughput
What happens if consistency exceeded?
get a 400 code
Web identity provider basics
Use Facebook/google or any open-id connect compatible to Id provider. Done with AssumeRoleWithWebIdentity API
Need to create role first
Parts of a Web Identity Provider link
Web identity token
App ID of provider
ARN of role
Temporary security credentials include
Access token/key
Expiration time limit
AssumeRoleId
SubjectFromWebIdentityToken
Steps taken to authenticate
User authenticate with ID provider
Passed a token by ID provider
Your code calls AssumeRoleWithWebIdentity API to provide token for the ARN of the role
You have a motion sensor which writes 600 items of data every minute. Each item consists of 5kb. What should you set the write throughput to?
50
Correct. 600 / 60 = 10 items per second. You are using 5kb per write and writes are measured in 1kb per write. This means that you need 5 writes per item, 5 x 10 = 50 writes per second.
What is the API call to retrieve multiple items from a DynamoDB table?
BatchGetItem
You have a motion sensor which writes 600 items of data every minute. Each item consists of 5kb. Your application uses strongly consistent reads. What should you set the read throughput to?
20
Correct! 600 / 60 = 10 items per second. 5kb rounded to nearest 4kb chunk is 8kb. 8 / 4 = 2. This means you need 2 reads per item. 2 x 10 = 20 reads per second. As the reads are Strongly consistent, you do not need to divide by 2. So the answer is 20.
You have an application that needs to read 25 items of 13kb in size per second. Your application uses eventually consistent reads. What should you set the read throughput to?
50
Correct. Your application reads 13kb per item. Round up to the nearest 4kb chunk = 16kb. 16kb / 4kb = 4 reads per item. You have 25 items so 25 x 4 = 100. Your application requires eventually consistent reads which means you divide by 2. 100/2 = 50.
What amazon service can you use in conjunction with SQS to fan out” SQS messages to multiple queues.”
SNS
