AWS B2B Data Interchange Cheat Sheet

• It manages the trading-partner lifecycle and automates the transformation and generation of Electronic Data Interchange (EDI) transactions, converting between standard EDI formats (such as X12, EDIFACT, HL7v2) and JSON/XML.
• It is fully managed (does not require an infrastructure to operate) and is designed for pay-as-you-go usage.

Key Concepts

• Profile = business identity in B2BI

• Transformer = translation/mapping logic (inbound/outbound)

• Trading Capability = defines documents & direction you support

• Partnership = connection with a trading partner + document types

Key Features

• Low-code trading-partner onboarding & mapping
  • Via a low-code UI, you can manage business partners, document types, and data flows
  • Generative-AI-assisted mapping helps create EDI-to-JSON/XML or JSON/XML-to-EDI mappings much faster
• Inbound & outbound transformations
  • Inbound: Receive EDI (X12, EDIFACT, etc.) –> convert to service-defined JSON/XML –> optionally map to custom format
  • Outbound: Accept JSON/XML or custom –> map –> produce EDI document for partner.
• Partner & workflow management
• Scalability & Integration

Example High-Level Architecture

Best Practices

1. Design modular mappings.
2. Collect representative sample documents for mapping.
3. Use the low-code UI and AI mapping to accelerate new partner setup.
4. Set alarms for failure or stuck transformations. Use logs and EventBridge to automate error handling.
5. Watch size limits for your S3 objects (e.g. max EDI file size is 150 MB, output will be 512 MB in some regions).
6. Track versions of transformer definitions by using versioning and change control.
7. Keep logs of transactions, acknowledge statuses, and partner change history.
8. Since many trading partners may be non-technical, provide onboarding guides and possibly managed SFTP/AS2 endpoints.

Security & Compliance

It leverages AWS shared-responsibility model: AWS manages infrastructure and you manage your data, buckets, IAM, encryption

• Supports S3 encryption (SSE-S3 or SSE-KMS) for input/output buckets.
• IAM integration: Use least privilege for rules/users managing B2BI resources.
• Generative AI mapping: your data is not used to train models.

Pricing

• No upfront costs
• You pay for active partnerships (each trading partner you maintain) and the number of transformation steps (each translation and mapping step)
• Note: Generative-AI-assisted mapping uses Amazon Bedrock behind the scenes. Additional Bedrock charges may apply for the mapping generation.

References

https://repost.aws/articles/ARvXPc–5rSC-We59JwQJwoQ/aws-b2b-data-interchange-introduces-generative-ai-assisted-edi-mappings

https://docs.aws.amazon.com/b2bi/latest/userguide/getting-started.html

https://docs.aws.amazon.com/b2bi/latest/userguide/what-is-b2bi.html