
<!DOCTYPE html><html><head>

<meta http-equiv="Content-Type" content="text/html; charset=utf-8">

  </head>

  <body>

    <div class="moz-cite-prefix">On 26/11/2025 10:32, David wrote:<br>

    </div>

    <blockquote type="cite" cite="mid:CAG-0zZdhzfX0m6bG5c0XGYTGuBBn2Vym6A06yZEpoPmz3=whfw@mail.gmail.com">

      

      <div dir="ltr"><br>

        <div>

          <blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-style:solid;border-left-color:rgb(204,204,204);padding-left:1ex">I

            assume it also benefits from being able to use HTTP

            persistent<span class="gmail-Apple-converted-space"> </span><br>

            connections.<br>

          </blockquote>

          <div><br>

          </div>

          <div>Just trying to understand, does that mean that virtual

            threads don't have this? Because running Socket Statistics

            in the background showing 17K established connections while

            running virtual threads vs 140 when using platform threads.</div>

          <div><br>

          </div>

          <div>At this point I am just trying to understand where this

            delay is coming from and why virtual threads trigger this.</div>

          <br>

        </div>

      </div>

    </blockquote>

    Persistent connections avoid needing a separate TCP connection for

    each request. LoomBenchmark's FixedPlatformPool appears to use a

    thread pool with 4 workers so it gets to benefit from the having the

    cached connection. Tomcat probably has more configuration to

    limit/configure the number connections to keep alive, probably 100

    or 200 by default. LoomBenchmark.java's virtualThreadExecutor is a

    completely different test. It attempts to establish 10k concurrent

    connections, it's probably causing connections to be closed and

    re-established at each iteration.<br>

    <br>

    -Alan<br>

  </body>

</html>