<div dir="ltr"><div>(Sorry, incomplete email; CTRL+Return didn&#39;t do what I was expecting...)</div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Mon, Dec 14, 2020 at 7:10 PM Dominique Devienne &lt;<a href="mailto:ddevienne@gmail.com">ddevienne@gmail.com</a>&gt; wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div dir="ltr"><div dir="ltr">On Wed, Nov 25, 2020 at 6:46 PM Niall Douglas via Boost-users &lt;<a href="mailto:boost-users@lists.boost.org" target="_blank">boost-users@lists.boost.org</a>&gt; wrote:<br></div><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">On 25/11/2020 10:54, Dominique Devienne via Boost-users wrote:<br>&gt; When I log the time it takes to resolve the server address and connect<br>
&gt; to it, then issue my HTTP request, it takes over 1 second (Win10,<br>
&gt; VS2019, C++17, Release, localhost for both client and server):<br>&gt; 2020-11-25T09:46:38.506180 Connected in 1.023s<br>&gt; <br>
&gt; While the same on Linux (RH7.5) is just over 2ms:<br>&gt; 2020-11-25T09:45:45.517550 Connected in 0.000s<br>&gt; <br>
&gt; That&#39;s a huge difference! Almost 500x...<br>
<br>
I can confirm that from libcurl wrapping ASIO, initiating a new HTTPS<br>
connection takes almost the same time whether from Linux or Windows.<br></blockquote></div></div></blockquote><div><br></div><div>I finally figured this out, mostly by chance...</div><div><br></div><div>As the extract below shows, Beast&#39;s example uses tcp::endpoint{net::ip::make_address(&quot;0.0.0.0&quot;), port}</div><div>in its sample usage, and I had copy/pasted that in my code (using &quot;0.0.0.0&quot;). This works, in that the server</div><div>starts fine, and clients work fine too, but leads to those &gt; 1,000ms client connection times on Windows.</div><div>While using e.g. tcp::endpoint{ tcp::v6(), port } results in 20-30ms connection times on Windows. (2-3ms on Linux).<br></div><div><br></div><div>So user error again. But maybe that usage example should use something else?</div><div>I guess taking an explicit IP is in case the server has multiple network interfaces / IPs?</div><div>Not sure what could replace it, to save the next poor soul from making the same mistake I did. --DD</div><div><br></div><div>PS: Niall, I&#39;ve still seeing 10x between Win10 and RH7, but that&#39;s better than 500x, and that&#39;s w/o SSL.</div><div>  Even with the server on Win10, and the client on Linux, it&#39;s still 3ms. So it&#39;s the Win10 client that&#39;s slow somehow I guess.</div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div dir="ltr"><div class="gmail_quote"><pre style="margin:1.5em;padding:0.5em;font-size:11.4px;border:1px solid rgb(204,204,204);overflow:auto;clear:left;color:rgb(0,0,0)">int main(int argc, char* argv[])
{
    // Check command line arguments.
    if (argc != 5)
    {
        std::cerr &lt;&lt; [...]
            &quot;    http-server-async 0.0.0.0 8080 . 1\n&quot;;
        return EXIT_FAILURE;
    }
    auto const address = net::ip::make_address(argv[1]);
[...]
    net::io_context ioc{threads};

    // Create and launch a listening port
    std::make_shared&lt;listener&gt;(
        ioc,
        tcp::endpoint{address, port},
        doc_root)-&gt;run();
[...]
    return EXIT_SUCCESS;
} </pre></div></div></blockquote></div></div>