AAI will follow very similar architecture to R2, and its position in the larger ONAP architecture remains consistent. Some additional microservices will be added, and we will make schema / edge rule changes in support of the various R3 use cases and functional requirements.
- the schema ingest library allows for a streamlined approach to using the AAI schema. Also allows for multiple OXM files so users can insert their own types at run-time.
Other updates:
AAI will use AAF for RBAC - currently planning on using Basic auth, since we haven't had success integrating with AAF to prove out a 2-way x509 cert exchange with AAF. We believe that this close to API freeze, forcing all clients to 2-way TLS is too steep a climb. We believe there also may be potential for integration with the proposed Pluggable Security Microservice, proposed by Andrew Baxter
New microservices:
enricher | Enables complementing AT&T data with federated data from additional sources. Exsiting seed code contributed from ECOMP |
cacher | The Response Caching Microservice (Cacher) is built to deliver multiple mechanisms of making API calls and populating the responses into a JSON datastore. Existing seed code from ECOMP |
validation | Microservice used to invoke validation mechanism. Used by POMBA, exsiting seed code from ECOMP |
AAI's schema/edge rules will not change for the Casablanca release to align with the modelling subcommittee's proposed information/data/runtime service and instance models. AAI is participating in the discussions and will map existing data objects to the approved clean versions when they are approved, targeting Dublin for potential changes.