Then there is the issue of having successful connectivity but miserable throughput. The throughput is what we need stats on. A website that I visit often hangs off a GigE switch on the hosting center. Traceroute shows 761 ms through the last hop which is the GigE switch. But pings succeed most of the time. Most SLAs that I have seen so far check for connectivity from usually a NOC to the customer, there is no mention of throughput anywhere. Bora ----- Original Message ----- From: "Sean Donelan" <sean@donelan.com> To: <nanog@merit.edu> Sent: Monday, December 04, 2000 10:57 AM Subject: Internet performance measures compared to other industires

I wonder if I'm sensitized to Internet problems in comparison to other industries. I checked a couple of other industries often given as examples of high uptimes, and found similar problems.

Ameritech will compensate 150,000 customers who waited more than 48 hours for a repair, or more than 7 days past the installation due date since January 1 2000. Ameritech is under investigation in four other states for missed dates and repair problems.

An e-merchant reports credit card fraud verification system nonfunctional 22 percent of the time. Mastercard and Visa didn't reply to press inquiries.

NASDAQ reports 99.96% system availability since January 1 2000.

AT&T, the only ISP I know publicly posting connection stats, reports 99% connect success last month.