Scaling - Dynamic DB

Question 1

How is DDB scaling measured?

Answer 1

Through throughput and a max item size of 400Kb of read/write units

Question 2

What is max traffic per partition?

Answer 2

3K RCUs or 1K WCU

Question 3

How does the partition key affect workload distribution?

Answer 3

The partition key port of the table’s primary key determines the logical location of data storage; input to an internal hash, output from the hash determines where the item will be stored

Question 4

What are features of auto scaling for DDB?

Answer 4

uses target tracking for target utilization; currently does not scale down well;

Question 5

What is on-demand DDB Scaling?

Answer 5

more expensive alternative to auto scaling, useful for unanticipated loads, instances allocated as needed,

Question 6

What is DDB Accelerator?

Answer 6

DAX is a HA, fully-managed, in-memory cache that sits in front of you DDB table

Question 7

What are the use cases for DAX?

Answer 7

require fastest reads, read intensive scenarios, repeated reads against DDB, bad for write intensive apps and apps that use client caching

Question 8

What is a scaling policy?

Answer 8

specifies read/write capacity min/max of a table or global secondary index that will be triggered