Hi Martin, On Thu, Jan 16, 2014 at 7:01 AM, Martin Sustrik <sustrik@xxxxxxxxxx> wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > On 15/01/14 21:05, Paul Colomiets wrote: > > > How does this work for with devices? I think it breaks the > > scalability principle. > > I don't think so. It's the same principle, but a different > granularity. With the req/rep as we have it today it's individual > requests that are load-balanced among workers. With "snappy" variant > it's clients that are load-balanced among workers. > > Of course, the latter it's somehow less scalable, but in an > environment with large amount of clients, each issuing negligible part > of the requests, it may work pretty well. > > From the text, it seems that only nearest peer id is checked against, right? So if we have the following scheme: (many clients) -> A -> B -> (many workers) Only a single worker is used, because all clients of A look like single client to B, right? -- Paul