AT&T; Corp. said Wednesday that software problems with one of its switches caused last week's outage of its nationwide frame-relay network for high-speed data services.
The No. 1 U.S. phone company said it has changed the procedure for upgrading switching software to prevent future outages. The frame-relay network, the largest of its kind in the U.S., shut down April 13, and some customers were without service as long as 26 hours.
Businesses, including airlines and banks, use the AT&T; network to send information between remote offices and automated-teller machines.
AT&T; Chairman and Chief Executive C. Michael Armstrong has been moving quickly to get to the root of the problem so he can assure customers it won't happen again. Since the outage, the New York-based company hasn't been charging for the service, which accounts for about $1 billion, or 2%, of AT&T;'s annual revenue and is increasing at 30% to 35% a year.
"This disruption certainly did not meet our customers' expectations for service reliability," Armstrong said in a statement. "We have taken steps to assure that this problem will not happen again."
Armstrong said he expects further investigation of the problem to be completed shortly and to resume charging customers for the service.
The outage started when one of the switches, used to direct traffic on the network, wasn't able to upgrade software and sent a large number of error messages that overloaded other switches in the network.
Cisco Systems Inc., which made the switches, worked closely with AT&T; to solve the problem. Armstrong said the problem didn't alter AT&T;'s confidence in Cisco and its products.
AT&T; shares slipped 6 cents to close at $63.31 on the New York Stock Exchange. San Jose-based Cisco fell 25 cents to close at $73.63 on Nasdaq.