Season 4 Episode 11: Dynamo DB Myths, Database Backups, and Leaky Buckets

In Season 4, Episode 11 Karl & Jon are joined by AWS Community Builder, Joe Stech. They discuss Amazon EC2 allowing AMIs now integrating with AWS Config, Amazon DynamoDB on-demand capacity mode, long-term backup options for Amazon RDS and Amazon Aurora, DeepSeek R1 model, misconfigured AWS S3 bucket exposing US nurses' data and then the guys started debating whether to call it "Glacier Potato" or "Deep Freeze Fries" as the next AWS storage tier!

 

04:00 - Amazon EC2 Allowed AMIs now integrates with AWS Config 

This feature allows easier monitoring of the impact of enabling allowed AMIs in EC2. It's particularly useful for regulated and secure environments where only approved, hardened images can be used. The integration with AWS Config simplifies the process of tracking and auditing AMI usage across accounts.

 

07:52 - Demystifying Amazon DynamoDB on-demand capacity mode 

The article addresses 11 myths about DynamoDB's on-demand capacity mode, covering cost, performance, scaling, and implementation misconceptions. The discussion highlights that many of these "myths" are not widely held beliefs among experienced users, but may be helpful for those less familiar with the service or dealing with outdated information.

 

19:00 - Long-term backup options for Amazon RDS and Amazon Aurora 

The article outlines various options for long-term database backups beyond the standard 35-day retention period. These include manual snapshots, using AWS Database Migration Service, exporting snapshots to S3, and database-specific dump tools. The discussion emphasized that while long-term backups are rarely used for recovery, they may be necessary for compliance and auditing purposes.

 

27:26 - DeepSeek-R1 now available as a fully managed serverless model in Amazon Bedrock

The DeepSeek R1 model is now available as a fully managed serverless model in Amazon Bedrock. This means users don't need to run the model themselves, and it's now priced per token like other managed models. The discussion touched on potential concerns about the model's Chinese origins and data security.

 

34:26 - Misconfigured AWS S3 Bucket Exposes Us Nurses' Data 

A misconfigured S3 bucket led to the exposure of sensitive data belonging to 86,000 US nurses. The discussion highlighted that while such incidents have become less common due to AWS's improved security measures, there might be a potential increase in similar incidents due to the rise of AI-assisted coding by less experienced developers.

 

Our guest's  blog: https://joeste.ch
and: https://learn.arm.com