Tim Brown wrote on 04/28/2015 09:16 AM:
Jay proved to me that the web-server was working with an experiment involving Apache Bench (ab); which is a nifty tool, I must say!
I also like JMeter for this purpose, since JMeter can validate output, for performance stress-testing. Though JMeter is more work to configure than `ab`, and its validation model isn't ideal.
(In some ways, would be nice to be able to use a Racket-based tool to do performance testing and validation. However, when you're performance testing a Racket-based server, you're also stressing parts of Racket that might also be stressed in your testing tool, and you don't want the same issue on both ends to, say, obscure the problem or the cause. Diversity is good. Also, presumably, tools like `ab` have been written to be pedal-to-the-metal.)
I'm still no wiser as to why the browser serialises GETs to a single server. Hypothetically, if I wanted to AJAX the replies from my script into two divs -- I need to know why it's taking me 2·n seconds, and not just n.
Google for "keepalive". Because of connection-establishing overhead, such as for TCP (and tons more atop TCP, if you're doing SSL). Also, a client making too many concurrent requests to the same server can be considered being a bad citizen. Imagine that every client of the server did that; server might need an order of magnitude more capacity, to avoid simply throttling or rejecting all of the connections during surges. And other factors.
BTW, the F12 button in your desktop Firefox might bring up a Firebug-like developer console that will, among other things, let you see a chart of the request timings for a page load. If it doesn't, install Firebug add-on.
Neil V. -- You received this message because you are subscribed to the Google Groups "Racket Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to racket-users+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
