From e1c438380124a29f71ee254cdf082cef85d4d33c Mon Sep 17 00:00:00 2001 From: Marco D'Agostini Date: Tue, 28 Jul 2020 20:36:56 -0500 Subject: [PATCH] new: DynamoDB --- DynamoDB.md | 20 ++++++++++++++++++++ 1 file changed, 20 insertions(+) create mode 100644 DynamoDB.md diff --git a/DynamoDB.md b/DynamoDB.md new file mode 100644 index 0000000..97186c2 --- /dev/null +++ b/DynamoDB.md @@ -0,0 +1,20 @@ +# DynamoDB + +It's a [[NoSQL]] [[Database]] from [[AWS]] + +## Differences from SQL Database + +- Records of a table can have different fields +- Tables can be partitioned using **partition keys**, I think you cannot query *fast* across different partitions. I think an analogy to **partition keys** would be like having a table across multiple databases in [[SQL]] Databases +- Tables also have **sort keys**, this determines how records are ordered, and you can only query contiguous records, I think you cannot filter nor change the order (maybe only reverse the default order) +- So basically **partition keys** and **sort keys** determines how the table can be queried. This limitations is the tradeoff that makes [[DynamoDB]] fast. +- There are some improvements with the connection management, I think it allows a lot more concurrent connections, but I don't know the details. + +## Use cases + +- High scalability +- Better integration with serverless architecture, because of better connection management + +## What led to learning this + +- Podcast: Full Stack Radio: 139: Alex DeBrie - DynamoDB for Relational Database Diehards