Integration considerations

  • 18 minutes

Throughout your upgrade journey, you need to be able to make thoughtful decisions about data integration scenarios when implementing finance and operations apps.

When choosing an integration pattern, you should consider several factors, as you would have when implementing Dynamics AX 2012. The difference is that, now, you are working on a cloud-based application within Azure.

Latency

Consider the following factors regarding latency:

  • Synchronous - Integration is triggered with an immediate response that is required.
  • Asynchronous - Integration is triggered with a delayed response that is required.
    • Near real-time - Minimal latency (<1 minute) is allowed between trigger and transmission.
    • Scheduled batch - Integration will occur on a scheduled basis with a pre-determined recurrence.

Message routing

Consider the following factors regarding message routing:

  • Point-to-Point - Dynamics 365 to Legacy System
  • Enterprise Service Bus - BizTalk Server and others
  • Broker, Hub & Spoke, or Extract, Transform, Load (ETL) - For example, SQL Server Integration Services

Frequency

The integration request can be classified into the following frequencies:

  • High - Seconds or minutes
  • Medium - Hours or days
  • Low - Weeks or months

Trigger

Consider the following factors regarding trigger:

  • On-Demand/Manual - Integration is manually initiated by an end user or an IT user.
  • Event Triggered - Integration is triggered based on an event or condition in the source or destination system.
  • Time/Date Scheduled - Integration is triggered based on a predetermined schedule.

Interaction/Operations

Consider the following factors regarding interaction/operations:

Create - Record will be created in the destination system. Read - Integration will query the source and return a specifically requested piece of data. Update - Integration will update an existing record in the destination system. Delete - Results in the deletion of a record in the destination system. Action -Integration will trigger a system event, for example, Calculate sales order price.

Batching

Consider the following factors regarding batching:

  • Unbatched - Individual records are sent in the integration request.
  • Batched - Records are consolidated for transmission through the integration request.

Volume

The number of records for each batched request should be either Low, Medium, or High volume.

Transport protocol

Consider the following factors regarding transport protocol:

  • Windows File Service and FTP/SFTP
  • HTTP/HTTPS, WCF Web Service, and SOAP Web Service
  • ODBC and Direct SQL Query

File formats and Schema/Data Dictionary

Consider the following factors regarding file formats and Schema/Data Dictionary:

  • File formats - Flat file, .NET TCP, XML, and so on
  • Message/API Schema - XSD / WSDL / OpenAPI (Swagger)

Data mapping/metadata definition

Consider the following factors regarding data mapping/metadata definition:

  • Canonical model - Integration is mapped to a standardized model in a broker or ESB.
  • Application specific format - Direct mapping between the source and destination systems.

Transformation and translation point

Consider the following factors regarding transformation and translation point:

  • ESB/Broker - Occurs in the broker system during the integration process (if no logic is required).
  • Destination - Occurs in the destination system (if business logic is required for the transformation).
  • Source - Occurs in the source system before the data is transmitted (if logic is required).

Error handling and reconciliation

Consider the following factors regarding error handling and reconciliation:

  • Notifications - Email, Dashboard, SMS, and so on
  • Reconciliation point - Application, Middleware, Monitoring utility, and so on
  • Reconciliation tools - Data quality services, reports, queries, and so on