32 lines (28 loc) · 1.16 KB

Dynamo DB

Type: Document Store
Serverless
Schemaless (like most document stores)

Components

Table - collection of data
Items - row/collection/document of one element w/i a table.
Attribute - keys in an item. The values can be 'scalar' or 'nested' like a JSON.

Concepts

Primary keys

Standard

Partition Key - key-value style lookups which returns an item
table defines a "hash attribute".
I think this probably uses consistant hashing underneath the hood.

Composite

Partition Key and Sort/Range Key
each document is partitioned by the partition key. Then each file is sequentially sorted based off of sort/range key.
Good for scans.
Combination of partition & sort keys must be unique

Secondary Index

Global secondary index - both partition and sort keys are different from table
Local secondary index - same partition key; different sort key

DynamoDB Streams

Each time an item is added/updated/deleted it will induce a change-capture
AWS Lambda can trigger off of change-capture.

Atomic Counters

An asynchronous way to increment to a counter. No need for an ack.
Good for use-cases where consistency barely matters.