AWS has launched DynamoDB for the entire world and it is an amazing piece of technology. Here are 8 features to get success by using DynamoDB:
1. Why do you really need DynamoDB?
If the right tool for the job is DynamoDB and you should be aware of it. If you require aggregations or possess a small amount of data or grained ability to combine lots of data together then DynamoDB is not the right choice. In such cases RDS or Aurora is the apt choice and where durability doesn’t matter Redis or ElastiCache is the right choice.
2. Know everything in detail about DynamoDB.
Although everybody reads the document there are few points that are missed like how to use the tool and laying out your data at scale. It is a pretty dense section. There are only few words about stress-testing as DynamoDB is not an open source.
3. For help ask Amazon
For checking the parts of the account AWS has lots of tools so do not worry. Everything from limit increases to detailed technical support, Amazon is always there for help. They are always helpful in getting us in touch with the right people and fast-tracking our support requirements.
4. Please read before you write
The write throughput is five times costlier when compared to the read throughput. If there are lot workloads towards writing then please check whether you can avoid updating it in place. Reading will help you to reduce your cost before writing as it will avoid lots of mistakes especially in a write-heavy environment.
5. Batch Partitioning and writing upstream
If the machine upstream in dynamo receives the key information then you can combine or group the data together and save writing on it. You can just write once per second or minute instead of writing every time you can group together all the information instead. You can manage your latency requirements with batching. Locking or race conditions can be avoided by Partitioning.
6. Throughput on spike and dynamic modification
By auto-scaling your DynamoDB you can get significant savings by a bursty traffic. By releasing the AWS feature you can learn more from the AWS blog. For extra cost savings, you can manage how DynamoDB throughput is offered vs how much is it in use with AWS Lambda and Cloud Watch events.
7. Make use of DynamoDB Streams
A not well-known feature DynamoDB can post all the changes to what is importantly a Kenesis requirement. For developing pipelines, streams are very useful and therefore you are not constantly Log all of your hot shards running SCANS or doing your own program.
8. Log all of your hot shards
While facing throttling error one must log particular key for update. Depending on how your data is laid out DynamoDB will perform differently. AWS engineers run DynamoDB as a cloud service. IT is definitely a great piece of technology. By using it correctly will help you earn more profit.
Stay connected to CRB Tech for more technical optimization and other updates and information.