[26107] in North American Network Operators' Group
Re: How to achieve application reliability
daemon@ATHENA.MIT.EDU (Sean Donelan)
Sun Dec 5 01:07:37 1999
Date: 4 Dec 1999 21:55:43 -0800
Message-ID: <19991205055543.23528.cpmta@c004.sfo.cp.net>
Content-Type: text/plain
Content-Disposition: inline
Mime-Version: 1.0
To: nanog@merit.edu
From: Sean Donelan <sean@donelan.com>
Errors-To: owner-nanog-outgoing@merit.edu
On Sat, 04 December 1999, James Smith wrote:
> internetsecure to type in the credit card. The problem with Round-Robin
> DNS is the possibility of the consumer's web browser picking up an IP
> address of a server that is down. If it was a real payment gateway, your
Finally, a problem I can agree with. Netscape's browser did some interesting
things for application reliability when accessing home.netscape.com. But for
other web sites it seems to be one strike and you're out. Other browsers
followed their lead. Actually, I think Mosiac was first, so the programmer
meme was already formed. The original CERN web browser did try alternate A
records. The CERN browser had a problem handling interrupts when the user
got tired of waiting, so the Mosaic "error-recovery" method of the user
clicking on refresh until it finally worked seemed like an improvement.
The law of unintended consequences?
The application programmers will say its the networks fault. The network
engineers will say its the applications fault. And the user says a pox
on all your houses.
