1. Integration Playbooks for Procurement
  2. Integration Methodology

Integration Methodology

Integrating Oracle Fusion Cloud Procurement with upstream or downstream third-party systems requires you to follow a comprehensive framework that helps you combine different systems or components into a cohesive whole.

This is an example of the integration steps you should follow.

1. Requirements Analysis

  • Document the systems involved in integration.
  • To serve business purposes, identify the data required for the integration.
  • Define the integration type (inbound, outbound, or process)
  • Identify data processing (real-time, near real-time, or batch)
  • Identify transaction data volume involved (low, medium, or high).
  • Identity any other nonfunctional requirements, such as data encryption.

2. Design and Development

  • Based on data required for integration, analyze business objects available in Oracle Cloud Procurement.
  • Identify the “source of truth” for each data entity.
    • Avoid moving data in both directions unless necessary.
    • Always create data first on the identified source of truth and connect it with other systems.
  • Before starting your journey, evaluate the previously built integrations available on Oracle Market Place.
  • For available business objects, identify the best integration option based on:
    • Data shape and frequency.
    • Volume (identifies the use of web services versus file import).
    • Performance considerations.
    • End-to-end automation support.
    • Best practices and constraints of each option.
  • Identify setup requirements:
    • Identify the authentication mechanism involved (basic authentication, OAuth, SAML, and so on).
    • Identify required roles for the integration user. Each business object needs a specific role before you can do CRUD operations on that object.
    • Server and network level configuration changes from external sources, for example, opening network ports and allow lists).
  • Identify custom mappings, transformations, or any lookups that are required. Identify where you'd be doing that transformation or custom mappings, whether at the source, in the integration layer (such as Oracle Integration Cloud [OIC]), or at the target.
  • Design for failure. Perform proper error handling and recovery. Be aware of the Oracle Cloud Procurement maintenance window.
  • Develop the integration solution according to the design specifications. This might involve coding, configuring existing tools, or a combination of both.
Note:

It's recommended to use JSON Web Tokens (JWTs) for authentication from the integration layer to Oracle Fusion, especially for high-volume flows.

Concurrent requests above the Oracle Identity Cloud Service (or Identity Domain) thresholds will cause requests to be rejected with HTTP-429 errors. For information about the thresholds for different tiers, see the API Rate Limits sections in the Administering Oracle Identity Cloud Service guide and in the Oracle Cloud Infrastructure Documentation.

3. Verification and Validation

  • Build a verification checklist.
  • Ensure the data is clean.
  • Ensure prerequisites are set.
  • Perform a test before migration using test data to cover all the use cases.
  • Conduct rigorous testing to ensure the integrated system meets all specified requirements.
  • Document each error and its resolution steps.
  • Verify that graceful handling of errors is in place when reviewing edge cases, such as when Oracle Fusion SaaS is down for maintenance.
  • During each quarterly update, make sure to check the verification list and run all test cases once the test pods have been migrated to the new release.

4. Production Transition and Monitoring

  • Prepare for the transition to the production environment. This includes completing deployment plans, setting up the production environment, and planning for data migration, if necessary.
  • If using OIC, ensure the integrations are exported from the test instance and deployed on production. For third-party integration platforms, ensure the same.
  • Replicate all the setups that have been done in the test environment in the production environment.
  • Do a smoke test for a single transaction before sending the entire load to production.
  • Have a strategy in place to recover transactions in case of an outage. Perform a mock drill of each outage scenario.
  • Continuously monitor integrations and check their health. Send notifications in case of errors. Generate reports of how many integrations fail daily and look for deviation from the standard.