Currently, for one to utilize request-response, they would have to implement it manually within a custom behaviour, which would be more low level, and likely create more complexities due to having to communicate with the behaviour. In this case, we could create simple request response functions, with requests being listened on through a stream with it supplying the channel to respond to the request while a separate function can send a request to a specific peer, or possibly to a set of peers.

Questions:

1. Do we have the requests and response be represented as bytes or should we have separate functions to serialize/deserialize the request/response?
2. What would be the upper size limit or should this be specific per peer? In other words, the limit set in the configuration during initialization.