At 02:16 PM 8/9/00 , Rick Jones wrote:
>> In general, portions of an HTTP request can be transmitted
>> over separate TCP packets if an application (in this case,
>> polygraph) sets the TCP_NODELAY option on the socket, and
>> if it writes portions of the HTTP request using multiple
>> calls to write() to the socket.
>I would probably claim that such an application was badly broken.
>Separate writes of logically associated data (say the headers of an HTTP
>request) is simply asking for trouble, and setting TCP_NODELAY is just
>an ugly kludge to try to cover it up.
In the scenario that we are talking about, a user wants to test
whether a cache device can handle HTTP requests that arrive across
multiple TCP segments. It would help if polygraph had a mechanism
that would enable testing of such a cache device.
The issue is not whether cache devices or webservers should
or should not do multiple writes. Rather the issue, as I understand
is that an intercepting cache device should be able to handle
HTTP requests that come in across multiple TCP segments. And
in order to test such a device, one could enhance the polygraph
>these opinions are mine, all mine; HP might not want them anyway... :)
>feel free to email, OR post, but please do NOT do BOTH...
>my email address is raj in the cup.hp.com domain...
This archive was generated by hypermail 2b29 : Tue Jul 10 2001 - 12:00:14 MDT