Modernizing Applications: From Relational Databases to DynamoDB Using AI-Guided Workflows

See how AWS makes database modernization real in this hands-on demo with Esteban Serna, Principal DynamoDB Specialist SA, AWS. We'll migrate a Typescript app from MySQL to DynamoDB using an AI-guided workflow. Along the way, we'll demonstrate how to use DynamoDB NoSQL Workbench for data modeling, AWS CLI for application refactoring and data processing, and NoSQL Workbench for data movement.

Chapters
0:00 – Intro and session setup
0:09 – Esteban introduction and role
0:38 – Agenda overview: complexity, workshop, pillars, solution, demo
1:00 – Why application modernization is complex
1:20 – Stakeholders: devs, architects, DBAs, business
2:17 – Anatomy of a typical current-state SQL application
3:05 – Application logic inside the database (SPs, triggers)
3:23 – Normalized schema structure in SQL
3:43 – What changes when moving to DynamoDB
4:29 – Schema transformation fundamentals
5:00 – Designing for access patterns and request volume
5:34 – Refactoring: moving from SQL to DynamoDB APIs
5:48 – Dual-write patterns & abstraction layers
6:22 – Why dual-writes matter for safe migrations
6:46 – Data size and downtime considerations
7:21 – Overview: the Database Modernizer Workflow
7:56 – What AI does vs what humans must do
8:40 – Modernization Pillar #1: Data Modeling
9:00 – Importance of understanding access patterns
9:24 – What the workshop analyzes in MySQL
9:59 – Reviewing application code & APIs
10:22 – Required artifacts: access patterns, velocity, sample data
10:49 – Pillar summary
10:54 – Modernization Pillar #2: Application Refactoring
11:03 – Implementing the abstraction layer
11:28 – Managing multiple access layers simultaneously
11:55 – Test-driven development and LLM hallucination control
12:56 – Implementing feature flags for migration safety
13:51 – Pillar summary
14:08 – Modernization Pillar #3: Data Movement
14:14 – Using dual-writes to minimize downtime
14:44 – ETL pipelines with Glue MCP / Data Processing MCP
15:03 – Why the three pillars apply to all modernization paths
15:21 – Solution architecture overview
15:47 – Environment setup: MySQL, backend, VPC, servers
16:22 – AI prompt structure: requirements, design, tasks
17:01 – Why prompts must be tightly scoped
17:54 – Workflow file structure (requirements, design, tasks)
18:45 – Overview of the Database Modernizer workflow
19:19 – Stage 1: Analyze source database
20:21 – Stage 2: Generate DynamoDB data model
21:51 – Migration contract.json: the most important artifact
22:40 – Stage 3: Create abstraction layer
23:23 – Stage 4: Implement DynamoDB logic
23:28 – Stage 5: Enable dual-writes & feature flags
23:42 – Stage 6: Deploy DynamoDB tables and Glue jobs
24:08 – Stage 7: Data migration & phased cutover
25:05 – Workshop portal and steps overview
25:46 – Demo: sample e-commerce app
26:29 – Walking through the workshop environment
27:07 – Stage 1 artifacts: access patterns, entities, logs
28:09 – Understanding the 48 access patterns
28:51 – Request-per-second breakdown
29:51 – Stage 2: Running the DynamoDB MCP server
31:02 – Providing traffic distribution to model RPS
32:52 – Validating source data and confirming requirements
33:57 – Generating the DynamoDB data model
35:02 – Reviewing and validating the generated model
36:56 – Requirements file deep dive
38:26 – Entity relationships & denormalization strategies
39:17 – Reviewing the proposed DynamoDB table designs
40:49 – Users table review and PK/SK strategy
42:17 – GSIs for user lookup (username/email)
43:23 – Product table review
44:05 – Category table review and optimization
45:03 – Orders and cart modeling
46:06 – Why related data should be colocated
47:08 – Providing model feedback to the MCP server
48:39 – Iterating on DynamoDB table design
49:42 – Fixing PK/SK, ID reuse, and index optimization
50:44 – Re-running and validating updates
52:17 – Reviewing updated data model
53:01 – Validating corrections and next iteration
54:03 – Workshop summary and next steps
54:38 – Closing remarks
55:00 – Live Q&A introduction
55:17 – Q1: Cost of migration & AI prompts
56:04 – Q2: Timeline estimates and OLAP modeling
57:38 – Q3: Do you still need DynamoDB expertise?
59:04 – Q4: How much time to spend on the workshop
59:49 – Q5: Running the DynamoDB MCP server
1:00:28 – Q6: Most important modernization pillar
1:00:43 – Q7: Data privacy considerations
1:01:14 – Closing Q&A

Видео Modernizing Applications: From Relational Databases to DynamoDB Using AI-Guided Workflows канала Eon_io