Domain Driven Design and Event Sourcing with RIM
Introduction RIMResistance is Futile, you Will be Assimilated!
Introduction DDD and Event Sourcing George de la Torre Developer of custom clinical apps For a long time Discovered RIM a long time ago… Oleg Sitalo (silent partner) Developer of custom clinical apps For a long time together Believes in RIM, at occasions…
Agenda DDD and Event Sourcing Google will find heaps of stuff on this Will focus on the RIM stuff Great resources for DDD/ES http://msdn.microsoft.com/en-us/library/jj554200.aspx Implementing Doman-Driven Design Vaughn Vernon Domain Driven Design (the Blue Book) Eric Evans
Domain Driven Design DDD and Event Sourcing Object Oriented Model Business needs & language captured HL7 v3 RIM benefits Design done! – the hardest part Standard – open & shared to all Deep insights into domain knowledge Integration of domains built in (No barriers!) Reuse logic across clinical domains
Domain Model Blueprint for all Aggregates HL7 V3 RIM
Supporting Domains External Services Vocabulary Bounded Contexts
Supporting Domains External Services Anti Corruption Layer Other systems or Big Ball of Mud… Business Process and Rules Engines Point: This layer keeps RIM pure!
Event Sourcing DDD and Event Sourcing Only the Aggregate (D-MIM) is persisted The command is saved (Behavior) Aggregates publishes the events Command Query Responsibility Segregation (CQRS) Command (Transaction) Model Query Model
Patient Dashboard Example Use Case Schedule a patient visit Find Patient (Query Model) View Calendar (Query Model) Patient assigned time slot (Command) CreateAppointmentCommand(data) executes Scheduling aggregate is created from data The created aggregate is appended Publishes event to update Query Model
Scheduling Aggregate Example
Scheduling Aggregate CMET (PAT) is the Root
Considerations Scheduling Aggregate Entity is usually the Root No cloning, only class codes Aggregates varies with the RIM set Act may be the Root (Strategy) Reference others Aggregates (Observations) Act Mood partitioning PHI segregation requirement (Query model too) Performance
Considerations continued… Scheduling Aggregate Considerations continued… Serialized, XML, JSON, etc… Relational Database (SQL) Cache based systems File systems NoSQL Natural for event stores Easier to query Aggregates if needed Aggregates loaded for logic on command
Scheduling Aggregate Appended CreateAppointmentCommand(data) Event Store ID Name Version Data 1 Schedule 1.2 2 3
Scheduling Aggregate Publishes Events AppointmentCreatedEvent(data) { UpdateDatabaseView() } Query model MRN First Last DOB Phone Start End Note
Scheduling Aggregate Publishes Events AppointmentCreatedEvent() { CreateAdmissionMessage() } HL7 2.x MSH|^~\&|ADT1|MCM|LABADT|MCM|198808181126|SECURITY|ADT^A01|MSG00001-|P|2.3 EVN|A01|198808181123 PID|||PATID1234^5^M11||JONES^WILLIAM^A^III||19610615|M-||C|1200 N ELM STREET^^GREENSBORO^NC^27401-1020|GL|(91-9)379-1212|(919)271- 3434||S||PATID12345001^2^M10|123456789|9-87654^NC
Scheduling Aggregate Publishes Events AppointmentCreatedEvent() { UpdateDataMart() }
Patient Dashboard Example Two Separate Models (CQRS) Domain (Transaction) Model Complex structure (RIM) Aggregates are only created from RIM Aggregates are appended (Write Only) Query Model Simplified model structure (SQL) Multiple models created as needed Model optimized for querying (Read Only)
Simple Component Design CQRS Web Dashboard View Model Aggregate Services App Services Publishes to Commands Query Events SQL