Cisco, IBM propose Internet-based disaster alert system

Engineers from Cisco Systems Inc. of San Jose, Calif., and IBM Corp. of Armonk, N.Y., have offered a set of basic guidelines for building an emergency alert system suited for the 21st century—one that uses the Internet.Such a system could be used to quickly alert people in the appropriate geographic area of an impending catastrophic event, such as a tsunami or a hurricane, said Fred Baker, a fellow at Cisco.Baker, along with Brian Carpenter, a senior engineer at IBM, submitted a of how such a system may work to the Internet Engineering Task Force on January 11. The two are now accepting feedback on refining the model. Although the idea of an Internet-based warning system has been floated for a number of years, Baker said he saw renewed interest after the December 26 tsunami ravaged southern Asia.Today, the government alerts citizens of natural disaster primarily through alerts sent to radio and television stations—courtesy of the U.S. Emergency Alert System—as well as through the use of sirens. “One problem … with simple sirens is that they communicate very little information beyond ‘Heads-up! Get more information,’” Baker said by e-mail. Existing warning centers, such as the NOAA’s or the Geological Survey’s , could send e-mail alerts using this approach, Baker said.By the engineers’ estimation, an Internet system would not require any new protocols, though some organization would need to coordinate efforts between government and private enterprises.Baker and Carpenter are not alone in this pursuit of designing a unified alert system. The Organization for the Advancement of Structured Information Standards, OASIS, has developed a vocabulary for alert distribution, called the . CAP defines a vocabulary for describing the severity and type of alerts.Eliot Christian, a data and information systems manager for the Geological Survey who helped developed the vocabulary, said that CAP complements Baker and Carpenter’s proposed system. An Internet-based alert system would deliver alert messages, but CAP could be used by communications companies to decipher and distribute the messages automatically. Baker offers a hypothetical example of how this system could alert people to an impending tsunami:A NOAA ocean buoy with sensing instruments could detect a series of large incoming waves. NOAA could coordinate this information with other warning centers, such as the Tsunami Information Centre operated by United Nations Educational, Scientific and Cultural Organization, to determine if the waves were part of a larger, more destructive phenomena. If so, NOAA could then send out a machine-readable alert, using the CAP protocol, to communications companies and emergency managers. These messages would describe the urgency, severity and certainty of the impending tsunami. To eliminate the chances of hackers spoofing alerts, an authenticated service of some sort would be essential, Baker said. A warning center could relay alerts over the Internet using Secure Multipurpose Internet Mail Exchange protocol, or S/MIME, messages that could then be relayed by cellular phone carriers over short message system protocol.A cellular telephone company could then broadcast an alert message to all of its phones in a geographic region, automatically making adjustments for the area’s language. “You’re asleep on the beach. Your mobile telephone starts beeping, indicating that it has received a text message. You open your bleary eyes, fumble with the thing, and read ‘Tsunami warning your region until noon. Move to a location above 100 meters. Further info +1-805-555-1212,’” Baker hypothesized. “Such an approach would, of course, not warn everyone,” the Internet draft stipulates. “Everyone does not carry a mobile telephone, and everyone who does would not necessarily read it. But it would warn a large percentage, which might help.”