Adaptive Architectures • Marty Pitt • GOTO 2023 (Category Architecture)

Interesting. report from Marty Pitt The same theme he had raised a year earlier in "Using Semantic Metadata to Create an Automated Microservice Data Mesh." told earlier. However, now this idea is fed through adaptive architecture, which is achieved by the fact that we do not write binding code. (glue code)We need to integrate our individual APIs to do useful work. As in the past, this is achieved by:

• Taxi languages for semantic markup of our existing APIs when we add metadata over it to describe semantics TaxiQL declarative query language to describe what we want to choose in terms of semantics
• Orbital Platforms to host these meta descriptions and execute these declarative requests, when under the hood the platform performs all requests and spits out the finished information

In summary, Taxi is similar to GraphQL in terms of the goals of providing a single entry point for combining multiple APIs. Taxi does not require resolvers, Taxi is protocol independent and can combine different APIs. (OpenAPI, RAML, JsonSchema and Protobuf)TaxiQL allows consumers to define their data contracts. But the problems are similar to those of GraphQL.

As a result, the adaptive architecture, according to the author, can be built on the basis of

• Taxi - semantic metadata for APIs
• TaxiQL - declarative query language
• Orbital - TaxiQL engine + observability

#Data #Software #SoftwareArchitecture #Architecture #Engineering