Imagine this scenario: Minutes before the recent Indian Ocean tsunami crashes into Sri Lanka or Thailand, the cell phones of people lying on the beaches there start to ring and a short text message issues a warning to flee to higher ground. Could a communications system like this have saved lives? Is it possible to build such a system by tapping GSM, a set of standards used globally by wireless service providers?
Imagine this scenario: Minutes before the recent Indian Ocean tsunami crashes into Sri Lanka or Thailand, the cell phones of people lying on the beaches there start to ring and a short text message issues a warning to flee to higher ground.
Could a communications system like this have saved lives? Is it possible to build such a system by tapping GSM, a set of standards used globally by wireless service providers? These are some of the questions that the Internet engineering community is asking itself in light of the recent tragedy.
"If you want to get people off the beach, you could send a [Short Message Service] to every cell phone that is active in a certain cell," says Fred Baker, a Cisco fellow who has expertise in Internet-based emergency communications. "You would know what cell phones were active in a particular cell near the coastline. You'd be able to target a particular geographic region by basing it on the GSM service."
Baker and other members of the IETF have been working for several years to create methods for prioritizing emergency communications over the Internet. The IETF is considering expanding this work in response to the devastating tsunamis.
"Distributed robust networks like the Internet provide a wonderful opportunity to spread information quickly and to a wide audience, thus assisting in disaster recovery," says Kimberly King, co-chair of the IETF's Internet Emergency Preparedness (IEPREP) working group and a principal network engineer at Science Applications International.
King says the Internet can help in all types of emergency communications - between the public and governmental authorities, from one governmental authority to another, and between members of the public trying to reach each other. "It is clear the Internet has much to offer each of these scenarios," she says.
However, the Internet has one major drawback when it comes to emergency communications: It's a best-effort communications system, so there's no guarantee that an important message will be received in a particular time frame. That's what the IETF is trying to fix with its ongoing development effort.
"With the tsunami situation, you would want an e-mail or instant message to go from the warning center to the right place and get there in a stated amount of time," Baker says. "The message must be known to be coming from an authenticated user, and it must be flagged as an important message that needs to get through in a certain amount of time. If the message wasn't able to be delivered in the stated amount of time, it needs to bounce back. That's the kind of capability we're talking about."
In contrast, the public switched telephone network (PSTN) can prioritize emergency communications for applications such as the U.S. Government Emergency Telecommunications Service (GETS), which is used in the aftermath of terrorist attacks, hurricanes and other natural disasters. With GETS, bandwidth is reserved for emergency calls, and users also can schedule calls to guarantee they go through.
The U.S. military also has special telephone systems that support pre-emption to make sure important calls - such as from the commander in chief - are connected regardless of other traffic on the system. Neither pre-emption nor prioritization of voice calls is possible over the Internet today using standard protocols.
"Today's VoIP services generally don't signal for bandwidth. They're fairly limited in their knowledge of the network," Baker says. "If we're going to do anything like call scheduling or call pre-emption in the Internet, we need a whole lot more intelligence in the network to make that happen."
So far, the IETF has not developed any new mechanisms or protocols for handling emergency communications. Instead, the group has been identifying requirements for Internet-based emergency communications and creating high-level frameworks that government agencies can use to create these systems.
The IETF formed IEPREP in response to a U.S. government request after the Sept. 11 attacks. Since its founding in February 2002, IEPREP has published four documents that outline requirements and terminology for emergency telecommunications systems. Three other documents, including a framework for supporting emergency telecommunications in IP telephony, await final approval.
"We were working on requirements and framework documents. To actually create an emergency telecommunications service, you really need to hone down to very specific scenarios," King says. "We tried to stay very general."
IEPREP plans to publish one more document, which will offer best practices for deploying Internet-based emergency telecom services using existing protocols. This document should be finished this year, King says. After that, the group's fate is unclear.
"IEPREP was only chartered to come up with requirements. It wasn't chartered to come up with any solutions," Baker says. "We're going to have to re-charter IEPREP to produce solutions or charter another group to produce solutions."
Another key piece of protocol work that's needed is a header for the Session Initiation Protocol (SIP) that can identify emergency communications and give them priority. This header could provide prioritization of emergency calls or military-style pre-emption.
"The header allows a caller to indicate that the call should have prioritized access to communication resources, such as IP telephony gateway circuits," says Henning Schulzrinne, one of the IETF's foremost authorities on SIP and a professor of computer science at Columbia University in New York. "It is used in conjunction with authorization."
Schulzrinne says the header, which the IETF's SIP group is working on, should be finalized in March.
Once the standards are done, the demand for Internet-based emergency communications systems is expected to be significant. The U.S. government has developed a prototype of a GETS-like service that would ride over the Internet instead of the PSTN. Meanwhile, the U.S. military and fellow NATO countries have told the network industry they need priority and pre-emption capabilities before they can migrate to VoIP.
In particular, the U.S. Defense Information Systems Agency is interested in the IETF's IEPREP work for its internal network. "They do have multilevel pre-emption and precedence currently on their circuit-switched networks, and they want to put it in their SIP and H.323 [videoconferencing] networks," King says.
The International Telecommunications Union also has a project looking at Telecommunications for Disaster Relief. This group is tracking the IETF's progress, King says.
"There is very strong governmental demand," Baker says. "I suspect that whatever the U.S. winds up doing will be adopted by the rest of the countries."
Some corporations with critical communications also might be interested in this capability for their private IP-based networks. For example, companies in transportation, energy or financial services might want to prioritize Internet communications in the case of a business emergency such as an airline crash, nuclear plant problem or stock market action.
"This is really only an issue where communication capacity is scarce," Schulzrinne says. "I could imagine using something like this for enterprises that have expensive and scarce circuits and may need these for high-priority business use...[like] oil companies having crews on 56K bit/sec lines in remote places."