On Sun, 2020-05-31 at 23:13 +0200, Michael Osipov wrote: > Am 2020-05-30 um 22:53 schrieb Oleg Kalnichevski: > > On Sat, 2020-05-30 at 21:58 +0200, Michael Osipov wrote: > > ... > > > > Now we need to decide if this is as good as it gets or we want > > > > to > > > > put > > > > more research into it. > > > > > > I read liked you idea of the chunks. 2048 seems too low for me > > > because > > > most well write with a buffer of at least 4 KiB. > > > > > > So I think we can try to things to miminize overhead: > > > > > > * Make this feature optional, just like we do with Expectation. > > > * Try to increase the step chunk size, e.g., 64 KiB or greater. > > > With > > > that we don't peek on each and every invocation. > > > > > > > Yes, we may consider making it optional / disabled by default > > Via RequestConfig? > > > We can give step chunking another go but it is less important now > > as I > > have inserted the check into the low level socket stream whose > > write > > operations should be already quite optimized by the protocol layer. > > But wouldn't that still reduce the check count because we'd hit the > wait > with SSLSocket? I am bit confused by your statement. >
It might but what I am trying to say is that even now there would likely be only one check on every large data chunk eliminating the worst case scenario of the application layer writing one or two bytes at a time each triggering a check and incurring a potential hit of 50 ms. Oleg --------------------------------------------------------------------- To unsubscribe, e-mail: httpclient-users-unsubscr...@hc.apache.org For additional commands, e-mail: httpclient-users-h...@hc.apache.org
