httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Rob Hartill <r...@imdb.com>
Subject Re: potential Apache performance problem: size of the TCP send buffer
Date Mon, 04 Nov 1996 21:23:48 GMT
Ben Laurie wrote:

>Brian Behlendorf wrote:
>> This patch looks good to me - a patch against the current cvs tree is appended.
>> Any other +1's?
>
>No doubt there'll be some portability issues, but +1 anyway.

Is the default set optimally now ?  This lets us change a value, but is
it set to a reasonable value by default ?

>> On Mon, 28 Oct 1996, John Heidemann wrote:
>> > Apache-1.1.1 doesn't appear to set the size of the TCP send buffer.
>> > On many systems the default send-buffer size is too small to support
>> > full-rate transfer over high bandwidth-delay-product networks (such as
>> > cross-US Internet traffic).  For example, a 100ms round-trip time
>> > (typical cross-US latency) and a 4kB default buffer size (common to
>> > many OSes such as SunOS) limit transfer rate to 40kB/s.
>> > 
>> > Much better cross-US transfer rates are possible between
>> > well-connected hosts if the send buffer size is increased.  A patch at
>> > <http://www.isi.edu/~johnh/SOFTWARE/APACHE/apache_1.1.1_isi_set_sndbuf.patch>
>> > implements the SendBufferSize directive to allow Apache to have a
>> > configurable buffer size.
>> > 
>> > A complete description of this directive is in the file
>> > send_buffer_size.html which accompanies the patch.  For a pointer to
>> > other patches, see <http://www.isi.edu/~johnh/SOFTWARE/APACHE>.  For
>> > descriptions of the problems motivating these patches, see
>> > <http://www.isi.edu/lsam/publications/phttp_tcp_interactions/>.  [For
>> > those of you who have seen this web page before, there are some new
>> > figures in ``The Performance Problems'' section and the buffer-size
>> > problem is described under ``Other Problems'' there.]
>> > 
>> >    -John Heidemann
>> >     USC/ISI

Mime
View raw message