Re: Preparations for ZD's upcoming Apache/Linux benchmark

Alex Belits (abelits@phobos.illtel.denver.co.us)
Wed, 9 Jun 1999 00:28:05 -0700 (PDT)


On Wed, 9 Jun 1999, Arjan van de Ven wrote:

> > There is another problem -- HTTP 1.1 (as opposed to HTTP 1.0 with
> > extensions) allows client to make multiple requests without waiting for
> > responses, then server should return all reaponses in sequence. I have no
> > idea, if anyone really supports that (can anyone show me an example that
> > does?),
>
> apt-get (Debian's package system" does. Squid does. In fact, I have seen my
> apt-get use it on my Squid.
>

Does Squid support it for its clients by itself, does it expect to
be supported by a server, or both? And if it supports this by itself, what
will it do in the case of DoS attack (send ahort request to something
slow-answering, then huge number of requests for something else in the
attempt to grow Squid in memory)?

[skipped]

> > of data while it sends back hundreds of small files over the
> > same high-latency connection back.
>
> It needs careful design. PUT and POST are for userspace anyway. I'll have to
> think about this some more..

Yes, but even if they are, they can be sent before previous request is
answered, so they can't be passed to userspace yet if the whole thing
should remain "transparent". Unless the responses are blocked
(causing major headache in another place -- with multiple requests,
received over the same connection, all sent to different "new" file
descriptors, so multiple processes will start answering, block on writing
and then kernel serializes them back in the order of requests).

-- 
Alex

---------------------------------------------------------------------- Excellent.. now give users the option to cut your hair you hippie! -- Anonymous Coward

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/