gaia and microservices on k8s about gaia

[romange]

gaia provides low-level rpc support with focus on performance and scalability. if you need 1M qps per node then gaia is better suited than grpc. See asio_fibers that demonstrates gaia rpc in action. grpc, on the other hand, has more features and is probably better suited for microservices. Google invested dozens of thousands of eng hours into grpc and it covers many languages. gaia rpc is currently c++ only. The open source version of gaia does not include protobuf bindings at the moment and high-level rpc protocol is not implemented.

It's still possible to use gaia to impelement resful backend however I did not design any high level classes for this. There is http_main binary that demonstrates how to register http handlers but again I did not invest much effort into designing a convenient RESTful framework. I am using http mainly for debugging and monitoring purposes.
There is also https client code that I recently wrote for connecting to GCS API.

To summarize, gaia can support everything you need but it's not out of the box. If you need to connect to k8s APIs you need to implement them yourself similarly to how I connected GCE/GCS.
if you need to develop a RESTful backend, you will need to think about what functionality is missing and implement it. Gaia will provide you with the framework to implement asynchronous computations and use multiple cores efficiently. If efficiency is not your priority than golang, python or java is better option :)

[romange]

FYI, currently, the only "industrial-grade" open-source descendent of GAIA is GAIA-MR that uses GAIA library underneath to expose high-level MapReduce framework.