Your team is developing an online bidding application which has used DynamoDB tables to store customers’ information. The application requires the fastest possible response time for reads in microseconds. However, it seems that the existing DynamoDB tables cannot provide such performance. Which approach is the best to improve the read performance greatly without modifying existing application logic?

Answer options:

A.Configure an on-demand read/write capacity mode for the DynamoDB table.
B.Create a Global Secondary Index for the DynamoDB table so that queries can be more efficient.
C.Create a read replica table in another region (Global Table) to improve the read capacity.
D.Configure a DynamoDB Accelerator (DAX) cluster for the application to use which can deliver up to a 10x performance improvement.

Correct Answer – D
DAX is a DynamoDB service to provide access to eventually consistent data with microsecond latency. Refer to https://docs.amazonaws.cn/en_us/amazondynamodb/latest/developerguide/DAX.html#DAX.use-cases
for its use cases.
Option A is incorrect: This option can scale up/down the read/write capacity however it cannot improve the performance dramatically.
Option B is incorrect: Global Secondary Index cannot guarantee that the performance will be improved. It is also up to how application code reads the DynamoDB table.
Option C is incorrect: This case does not mention that multi-region table is required. It cannot improve the performance if the application only runs in a region.
Option D is CORRECT: DAX is a fully managed, highly available, in-memory cache that has a 10x performance improvement for DynamoDB. Besides, as DAX is compatible with existing DynamoDB API calls, there is no need to modify application code for this approach.