Avoiding Telephone Woes

July 05, 1991

Computer glitches which knocked 5 million Chesapeake & Potomac customers off the line last week, repeated in California and Pennsylvania, have sparked new questions about telephone reliability. Just as the U. S. Telephone Association was lobbying Congress for new laws to let phone companies into new lines of business, the lines serving regular phone service were locking up.

A faulty circuit board is the chief suspect. The board was not installed in the regular telephone switch, but in Signaling System 7, a sophisticated external network which directs traffic between telephone exchanges. The system is supposed to speed calls; where old-style networks connect voice circuits for every call, even if the called line turns up busy, System 7 avoids bottlenecks by checking the proposed route first, then ordering connections when all's clear. It's still new, but it's faster and it makes possible many new services.

Last Wednesday, however, System 7 computers in Baltimore, Hyattsville, Washington and Richmond went down, blocking linkages between local exchanges. Calls moving through a single switch could still get through, but transfers to another switch failed. That brought immediate questions whether the telephone companies were adept enough to handle the new technology as well as the new businesses they want to start. As one regulator put it, they should tend to their own knitting, providing plain old telephone service.

But plain old telephone service has grown complicated. U.S. customers now use 200 million phones, a third of all the world's phones. Ballooning demands have added new area codes to every major metropolitan area. New integrated services digital networks, cellular phones and improved paging and messaging systems load on ever-increasing traffic.

Federal regulators have called for more backups, but it is unclear where these would fit in. Signaling System 7, like all

computer-oriented switching circuits, uses duplicate processors at each site. A proper checking system should have shut down the processor with the bad board, allowing its local backup to take over. Moreover, the three working traffic computers, electronically paired and sharing traffic-cop duties via a special control-transfer "algorithm," should have been able to recognize the Baltimore failure and to ignore the faulty computer last week, limiting the outage. That didn't happen, prompting questions about the processors and their software.

Telephone engineers across the country are interested in this one, since it has triggered phone brownouts in switches made by more than one vendor. It's likely they will solve this particular problem, however, for it occurs only in one company's computers. In the meantime, there's no going backward. The computer revolution and the AT&T breakup and its aftermath changed things forever.

Baltimore Sun Articles
Please note the green-lined linked article text has been applied commercially without any involvement from our newsroom editors, reporters or any other editorial staff.