Discovery automatically assigns a module to an Architecture Canvas layer, following the module ' s naming convention.

Discovery only validates architecture rules at the module level.

Discovery allows you to determine for every module, which elements are being consumed by other modules.

Discovery cannot categorise modules to domains

Screens as layouts that have placeholders and load content from blocks. The blocks come from CW modules.

Include the Common flow and Menu artifacts

Put CSS in blocks or screens. Allows better performance and seamless screen transitions

Blocks grouped in functional modules with independent life cycles

CSS should only be placed in the mobile theme module

Poor service abstraction

Unmanageable dependencies

Manages complexity

Slow-moving legacy systems

User Experience Expectations

Integration Technology

User Stories, Personas and Roles

Architecture Canvas Modules

Information Architecture

All public entities in Core Modules should be set to Read-only

No business logic in the Foundational Modules

No core entities in Foundational Modules

No screens in the End User Layer

No front-end screens in Core Modules

Same as ECS with local replica but synchronization logic is separated. Pro: Code ^ independence. Consumers of CS is not affected by Sync. Sync can orchestrate several CS

... Entity is exposed as read-only and API is available to centralize business logic for entity creation/update

... a wrapper used to contain the logic, actions and data that will expose code that is inside of O external library or to inspect external database and import the data structures so they can be used as entities inside of OS

Same as Base ECS pattern, but have a local replica. Store data to serve as a local cache. Pro: Leverage Entity Use, Simpler Integration API. Con: Less impact on source system

... caches only summary data that is frequently lister, joined or searched. Full detail for a O single entry is fetched directly from external system. Use when whole database too big or costly to synchronize. Details are only required for single entities (not lists)

... Entity is not in Outsystems but in an external ERP system. IS just makes remote call to p external system/database. No data is being kept inside OS. Data retrieval may not be optimized as it needs to traverse two different systems to get the information back. Con: Integration API must support all use cases

... tries to fetch data from local cache entity, if not there, get single entry from the external v system. Cache only that record (read-through caching) Use when whole database too big or costly to synchronize. Integration only touches a small portion of the database. Avoid if access to lists of data is needed up front

Optimizes lifecycle independence

Promotes microservices

Minimizes impact of changes

Promotes abstraction of reusable services

Stop unmanageable dependencies

Prevent slow deployments, inflexible legacy systems

Avoid poor service abstraction