This is slightly changed in http 2.0 (unsolicited response) On Thu, Aug 17, 2017 at 6:24 PM, Todd Montgomery <[email protected]> wrote:
> REST and RPC have good uses. But they are not a panacea. > > REST has implicit coupling because each HTTP request must have a > corresponding HTTP response. Think of this as mandatory coupling from the > protocol perspective. This implies that management between a request and > its response must be done. (For retry, error, etc.). Whether you wrap that > in a future or not is somewhat cosmetic. Web Services and the protocols > attached to them are designed to try, as best they can, to decouple > requests/processing/response.... but can't separate request and response > from the HTTP perspective. > > From a higher perspective, HTTP can be made to look asynchronous from the > request, response, and processing angles.... at the application level. > However, most of the time, this is rarely done. Or done well. > > To allow full async communication, the basic communication block must be > simplex (1-way) and not have response dependency at the application nor > protocol level. > > This may seem that we are bashing on REST... and we are... somewhat. > Mostly, though, we are bashing how REST is used. More like a hammer and > everything is a nail. While in many cases, not nearly everything is a nail. > And the hammer leaves a lot of "collateral damage" in the form of coupling > and dependency that is unnecessary. For example, an extremely common > problem with some systems using REST is throughput being restricted by RTT > due to response coupling (application or protocol). > > On Wed, Aug 16, 2017 at 10:19 PM, Ziad Hatahet <[email protected]> wrote: > >> Martin mentions at 30:22 during his talk[1] that REST, while still better >> than RPC, still causes coupling between communicating actors as opposed to >> messaging. >> >> My question is: what's preventing REST calls from being wrapped in a >> future or higher level messaging construct (e.g. Akka actor messages) such >> that communication is decoupled because it becomes asynchronous? >> >> >> [1] Beginning of discussion at https://www.youtube.com/watch? >> v=oxjT7veKi9c&t=1805 >> >> Thanks, >> >> -- >> Ziad >> >> -- >> You received this message because you are subscribed to the Google Groups >> "mechanical-sympathy" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to [email protected]. >> For more options, visit https://groups.google.com/d/optout. >> > > -- > You received this message because you are subscribed to the Google Groups > "mechanical-sympathy" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > For more options, visit https://groups.google.com/d/optout. > -- Studying for the Turing test -- You received this message because you are subscribed to the Google Groups "mechanical-sympathy" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
