Best server-less NoSQL database for key value retrieval
April 30, 2019
Best server-less NoSQL database for key value retrieval
Score 9 out of 10
Vetted Review
Verified User
Overall Satisfaction with Amazon DynamoDB
In our organization, we are moving towards server-less for our in-house application as well as suggesting the same to the client. Dynamo being a serverless NoSQL database plays a key role in our application architecture.
We are using Dynamo widely for session management and product catalog user profiles.
We are using Dynamo widely for session management and product catalog user profiles.
- Pay as you go, no upfront payment
- You can scale down your table to 1wcu 1rcu which bring its cost to a few dollars a month.
- Configuring auto-scaling
- Support to burst traffic
- Database as an API, so it can be easily integrated with the application.
- Wide support with other AWS services.
- A limited number of primary and secondary indexes.
- Poor performance for batch retrieval.
- Very bad support for operation where the aggregation is needed (sum, avg, min, max).
- No Joins, (it's not made for join).
- Saved lot of money for us because of its support to auto-scaling.
- As its NOSQL so if you are not doing proper analysis before deciding to go with NoSQL it can be disastrous, fortunately, we never faced such negatives.
When we were selecting for a server-less NoSQL solution, we had two option either to go with Cosmos or with Dynamo. Cosmos has a few great features, but in Cosmos, we need to provision at least 400 RU (request unit), which will make a single table cost at least 24$/month. While in Dynamo we can even provision 1 WCU RCU which makes it very economical. Cosmos has many great features as compared to Dynamo but in most of the use cases, we didn't require those so we went with Dynamo.