7 Years of DDD: Tackling Complexity in Large-Scale Marketing Systems
summary
Plexop is a large marketing platform spanning a dozen domains (campaigns, publishers, creatives, leads, commissions) built by a tiny team under aggressive deadlines that repeatedly shifts between monolith, bounded contexts and microservices. Key claims: ubiquitous language prevents costly inconsistencies; use four implementation patterns (transaction script, active record, domain model, event sourcing) guided by heuristics. Concrete tools: ElasticSearch, S3 projections, AWS ML, events stored in DynamoDB (leads ≈500 events). Pitfalls: stored procedures and over-granular services create a distributed monolith; rewrites of lead and bonus systems succeed after refactoring.
AI-written summary · about
chapters (24)
- 0:00Introduction
- 0:52Presentation Structure
- 1:33Vladiks Background
- 3:02Implementing the first third of our value chain
- 3:38Eric Evans seminal book
- 4:34Architecture
- 12:07The Problem
- 12:57Conways Law
- 13:21A Nightmare
- 13:57Domain Models
- 14:34Core Knowledge
- 17:14Commissions
- 21:05Market Hub
- 24:12Accidental Complexity
- 25:02The Bounded Context
- 28:24Business Domains
- 31:13Changing Domains
- 33:56Benefits of Changing Domains
- 35:19Choosing Patterns
- 42:43Bonnet Context Internals
- 43:39Suicidal Boundaries
- 46:23Trial and Error
- 50:58How DDD survived
- 52:04Questions
related talks
25mThe modern DXP
45mLearning to Love DDD - a Tale of Two Products
23mDesigning Robust and Scalable Distributed Applications
16mYou Can't Outrun Complexity - But You Can Orchestrate It.
30mMonolith to Microservices
23mMarketplace by APIs: Product Thinking, Integration Patterns & Platform Evolution
23mMicroservices within a Monolith
31mHow to set up your organization and architecture for growth?