DynamoDB: Under the hood, managing throughput, advanced design patterns | Jason Hunter | AWS Events

A deep dive into DynamoDB looking under the hood and showing advanced design patterns including: partition management, multi-AZ durability, read and write limits at the table and partition levels, On-Demand and Provisioned capacity modes, schema designs, and high-scale usage patterns. Part 2 of 2. Presented by Jason Hunter at the Database Modernization Week event. Part 1 available at • DynamoDB: Its purpose,... .
Learn more about DynamoDB at: go.aws/3Ra4C5f
Subscribe:
More AWS videos bit.ly/2O3zS75
More AWS events videos bit.ly/316g9t4
ABOUT AWS
Amazon Web Services (AWS) hosts events, both online and in-person, bringing the cloud computing community together to connect, collaborate, and learn from AWS experts.
AWS is the world’s most comprehensive and broadly adopted cloud platform, offering over 200 fully featured services from data centers globally. Millions of customers-including the fastest-growing startups, largest enterprises, and leading government agencies-are using AWS to lower costs, become more agile, and innovate faster.
#AWS #AmazonWebServices #CloudComputing

[@benlorence7390]

This is the best DynamoDB video I've seen - well done!

[@jasonhunter1786]

Fun fact: After recording this video I realized I spoke one inconsequential factual error. Being a perfectionist, this bugs me. But it was too late to re-record. What it means is you all have an Easter Egg to find. Let me know if you find it.

[@user-hs8hq8fv2p]

Hi Jason, is this about the read replicas you mentioned while talking about the DAX?

[@vadimshun726]

Thanks Jason, Helpful refresher, like Under the Hood concepts in general.

[@golamkabir99]

this is a great video on DynamoDB - thanks Jason

[@tranminhhaifet]

Thank you, very clear

[@carnelyve866]

Excellent!!!!

[@amazonwebservices]

Thanks, Carnel! ☁️ 🤝 😊

[@badscrew4023]

Good talk!

[@_root]

Thanks for great presentation.
On 70k read case of hot item - what would be a solution if you need to have a strong read consistency? DAX would not fit in this case.

[@jasonhunter1786]

You can create multiple GSIs with the same keys and schema as the base table and have your reads randomly pick where to read, as a way to scale reads super high at the cost of more writes and storage. It doesn't get you exactly strong read consistency but it's very close since GSIs usually update extremely quickly.

[@sherrywei9263]

Jason, great video, learned a lot as a new comber. Did the mistake happen in the end when you say throughput cost is $1 and storage cost is 50c, you should consider switch to standard-IA? It should be the other way around, i.e., if storage cost is $1 and throughput is 50c, you should consider switch to standard-IA. did I get it right?

[@jasonhunter1786]

Actually I said it right. The storage discount is 60% and the throughput uplift is 25%, so the crossover point is around when storage is 42% of throughput.