robiwano/siesta

question about the surveyor pattern

Opened this issue · 2 comments

Deleted user commented

Hi @robiwano ,

Hope you are all well !

I was wondering if it is complicated to extend siesta with the surveyor pattern and replicate what is described in this blog post https://daniel-j-h.github.io/post/distributed-search-nanomsg-bond/ ?

The idea would be to configure rest routes binding several, probably dockerized, respondent clients and aggregate all responses into one. Maybe, using a yaml file to configure the distribution tree ?!

Do you think it is doable ?

Thanks for your inputs and insights on that.

Cheers,
X

robiwano commented

Although I'll have to understand what the pattern does, I'm not sure I'd like to integrate that kind of functionality directly into siesta, since the aim is to keep it as simple as possible, but not simpler.
That being said, since siesta can create/remove routes dynamically, I think you'd be able to build something similar upon siesta.

Deleted user commented

Hi,

Indeed, the dynamic routes features triggered my interest.

https://github.com/daniel-j-h/DistributedSearch is an implementation of the article.

As much more a gopher than c++ developer, that's why I asked the question.

I am really interested by this feature for some deep learning rest api inference aggregation. There is a real use case ^^