Keynote/Boardwatch Results

James R Grinter jrg at blodwen.demon.co.uk
Wed Jul 16 22:34:09 UTC 1997


c-huegen at quadrunner.com (Craig A. Huegen) writes:
> It's a rough measurement, and if you'd go so far as to assign a 20% error
> margin, you'd stillsee that a web server still owns a *significant* piece
> of the click-to-data time, over 50%.

Especially if it's using the particularly sub-optimal (aka 'broken')
network stack that a very popular server operating system has.

In fact, some recent measurements of mine show a large variance even
for a connection setup on a local network depending upon what IP
stacks are involved: varying between 0.39s and 0.007s. (Unsurprisingly
the broken stack referred to above works quite well with itself, and
not too bad with an earlier OS from the same company - if I didn't
know better I might believe that they only tested with their own
systems.)

(this isn't really operational so if you want names let me know
privately. Perhaps someone out there has some clout.)

James.



More information about the NANOG mailing list