It''s not about scalability on a huge site, but the lowest possible
latency for a very small app.
I have a little app with Thin and Sinatra. My app generates the page in
about 50ms. After getting the browser request, thin waits 50ms and then
it takes 300ms sending the page, to Chromium at localhost.
(It takes 300ms to send a static a.html with the word "test" on it,
too,
so it seems to be a lower bound on my machine)
Why is that? Should Mongrel be faster sending the data? Or maybe
Passenger?
The generating time was taken on Sinatra, by doing Time.new.to_f then
printing the diff. The wait / send time was taken at Chromium
"resources" tab.
The fact that thin waits 50ms before starting to send anything makes
sense, if the app happen to set the header at the end of the output.
But spending 300ms just _sending_ the page to localhost page does
not. It''s a 1kb file. Maybe it''s forking before every page?
(But it
doesn''t reload automatically like shotgun, so this makes no sense; it
must be something else)
I could still optimize my app (since its performance is bound by an sql
query that is taking less than 10ms), but my bottleneck isn''t there,
it''s on Thin.
For comparison, djb''s publicfile can send a 100kb image in this same
300ms. (I know it''s unfair)
Some versions: ruby 1.9.2p0, thin 1.2.7
--
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Talk" group.
To post to this group, send email to
rubyonrails-talk-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To unsubscribe from this group, send email to
rubyonrails-talk+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
For more options, visit this group at
http://groups.google.com/group/rubyonrails-talk?hl=en.