dynamodb.md

DynamoDB

• Fully managed cloud database

• Seamless on-demand scaling

• Unlimited concurrent Read/Write operations

• Single-digit milisecond latency (< 10 ms)

• Sub-microsecond latency with DAX

DynamoDB components

• Tables: consists of a set of data
• Item: a table consists of multiple item, alike key-value pair
• Attributes: an item in a table consists of multiple attributes

Primary key has 2 types:

• Partion key (hash key) => simple primary key: which determines in which partition the data is stored. Such as: employee_id
• Partition key and sort key (range key) => composite primary key: determines the place of an item in a partition
• DynamoDB stores all items with the same partion key together. In a partition, items are order based on a sort key value

Secondary indexes in a table

• It is an alternate way to query table data using the primary key
• Consists of 2 type:
  • Global secondary indexes (GSI): partition key and sort key different from the primary keys defined on the table
  • Local secondary index (LSI): the same partition key as of the table but uses a different sort key
  • Limit 5 GSI and 5 LSI on a table

Dynamodb tables

• Top level entities

• Independent entities

• Flexible schema

• Control performance at the table level with RCU and WCU

DynamoDB naming conventions

• Prefix table name to create namespaces

• prefix.tablename or prefix_tablename such as: test.users, test_projects

• Best practice

Types of dynamodb

• String, number, binary, boolean (true/false), null (unknown or undefined state)

• Set such as ["red", "blue", "green"]

• Document (List, Map) such as: ["John", 22, "male"]

{
    "username": "example",
    "password": "pwd"
}

Table index

• Mandatory primary key => either simple or composite

• Simple primary key => only partion or hash key

• Composite primary key => partition key + sort or range key

• Partition or hash key decides the target partition

DynamoDB stream, trigger when database changed (update, delete, add)

Setup read consistency model with ConsistentRead parameter

• Set true if DynamoDB uses strongly consistent reads
• Set false if DynamoDB uses eventually consistent reads

DynamoDB partition and data distribution

• Table partitioning is a mechanism to segregate a large table into smaller, more manageable parts without creating a separate table for each part.
• DynamoDB manages data in partitions and uses SSD to storing data and automatically replicates data across multiple AZs in an AWS region

Data distribution with partition key and sort key

• Using partition key for determining target partition, where data is stored
• DynamoDB orders the data by the sort key value

Note:

• GSIs are also segregated in partitions
• The data in an index is stored separately from the base data of a table
• Index partitions and table partitions act in a similar manner in DynamoDB
• The best way to optimize partitioning is to use more distint values such as: COFFEE01, COFFEE02, ...

Secondary index

• Main purpose is to provide an alternative key for query operations
• The secondary index is linked with a table, which becomes the source of data for it
• When you change anything on the base table, the change automatically reflects into indexs such as: add, modify, delete item in the table

Global secondary index (GSI)

• Different partition key and sort key from base table
• When any request is executed against a GSI, its scope spans across all the partition in a table
• Simple attribute primary key or composite key
• There is no restriction on the size of the index
• Only support eventual consistency reads

Local secondary index (LSI)

• The same partition key as its base table => mapping between index and base table. It has a different sort key
• Only can composite key and query a single partition with a partition key value specified in the query
• Each partion size < 10GB
• Support both strongly and eventual consistency reads

Integration

For example, connecting DynamoDB by aws-sdk

const AWS = require("aws-sdk");
AWS.config.update({ region: 'us-west-2' });

const dynamodb = new AWS.DynamoDB();

dynamodb.listTables((err, data)=>{
    if(err) {
        console.log(err);
    } else {
        console.log(data);
    }
});