Internet Engineering Task Force
Internet Draft                                            H. Schulzrinne
                                                             Columbia U.
draft-schulzrinne-sipping-sos-04.txt
January 8, 2003
Expires: June 2003


     Emergency Services for Internet Telephony based on the Session
                       Initiation Protocol (SIP)

STATUS OF THIS MEMO

   This document is an Internet-Draft and is in full conformance with
   all provisions of Section 10 of RFC2026.

   Internet-Drafts are working documents of the Internet Engineering
   Task Force (IETF), its areas, and its working groups.  Note that
   other groups may also distribute working documents as Internet-
   Drafts.

   Internet-Drafts are draft documents valid for a maximum of six months
   and may be updated, replaced, or obsoleted by other documents at any
   time.  It is inappropriate to use Internet-Drafts as reference
   material or to cite them other than as "work in progress".

   The list of current Internet-Drafts can be accessed at
   http://www.ietf.org/ietf/1id-abstracts.txt

   To view the list Internet-Draft Shadow Directories, see
   http://www.ietf.org/shadow.html.

Abstract

   This document describes how Session Initiation Protocol (SIP) user
   agents and proxies can set up emergency calls and, more generally,
   reach emergency assistance via SIP requests. For that purpose, it
   defines a universal emergency SIP URI, sip:sos@domain and
   sips:sos@domain, that allows SIP user agents to contact the local
   emergency number. It also defines conventions that increase the high
   probability of reaching the appropriate emergency call center. The
   document does not define any SIP protocol extensions.



1 Introduction

   Using the PSTN, emergency help can often be summoned at a designated,



H. Schulzrinne                                                [Page 1]


Internet Draft                    SOS                    January 8, 2003


   widely known number, regardless of where the telephone was purchased.
   However, this number differs between localities, even though it is
   often the same for a country or region (such as many countries in the
   European Union). For end systems based on the Session Initiation
   Protocol (SIP) [1], it is desirable to have a universal identifier,
   independent of location, to simplify the user experience and to allow
   the device to perform appropriate processing. Here, we define a
   common user identifier, "sos", as the contact mechanism for emergency
   assistance. This identifier is meant to be used in addition to any
   local emergency numbers.

   We also describe how emergency calls are routed to the appropriate
   emergency call center (ECC). (In the United States and Canada,
   emergency call centers are referred to as Public Safety Answering
   Points (PSAPs).) Since each emergency call center is generally only
   responsible for a specific geographic area, it is important that
   calls are routed to the correct ECC. Regardless of whether the ECC is
   connected to the PSTN or is directly reachable via SIP, the network
   location of the caller has little relationship to its physical
   location. If the call is routed through a PSTN gateway, the
   originating number is likely either associated with the gateway or is
   permanently assigned to the IP phone, regardless of where it is
   currently located. For SIP-based ECCs, the IP address or Contact
   header information in the call only provides crude approximation as
   to the geographic location of the caller and may well be completely
   wrong if virtual private networks are used. Thus, the SIP request
   needs to convey the location of the caller so that the call can be
   routed appropriately. Section 6 discusses one possible approach.

   This document does not introduce any new SIP header fields, request
   methods, status codes, message bodies, or events. User agents unaware
   of the recommendations in this draft can place emergency calls, but
   may not be able to provide the same user interface functionality. The
   document suggests behavior for proxy servers, in particular outbound
   proxy servers.

1.1 Terminology

   In this document, the key words "MUST", "MUST NOT", "REQUIRED",
   "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY",
   and "OPTIONAL" are to be interpreted as described in BCP 14, RFC 2119
   [2] and indicate requirement levels for compliant SIP
   implementations.

2 Requirements

        o SIP-based end systems must be able to reach emergency call
          centers. These emergency call centers may have SIP



H. Schulzrinne                                                [Page 2]


Internet Draft                    SOS                    January 8, 2003


          capabilities or may be reachable only via a SIP-to-PSTN
          gateway. Since each ECC serves only a limited geographic area,
          often defined by jurisdictional boundaries such as state,
          province or county, SIP-based emergency requests must

        o While current emergency call centers are limited to voice and
          TDD (telecommunication device for the deaf) communications,
          future SIP-based ECCs should handle all relevant means of
          interaction, including multimedia and instant messaging [8].

        o It should be possible for devices to provide user interfaces
          that can directly cause an emergency call, without the user
          having to "dial" or type a specific address.

        o Even as each country is likely to operate their emergency
          calling infrastructure differently, SIP devices should be able
          to reach emergency help and, if possible, be located in any
          country.

        o While traveling, users must be able to use their familiar
          "home" emergency identifier. Users should also be able to dial
          the local emergency number in the country they are visiting.

        o Any mechanism must be deployable incrementally and work even
          if not all SIP entities support emergency calling. User agents
          conforming to the SIP specification [1], but unaware of this
          document, must be able to place emergency calls, possibly with
          restricted functionality.

        o Given incremental deployment, emergency call functionality
          should be testable by the user without causing an emergency
          response.

        o Emergency calling mechanisms must support existing emergency
          call centers based on circuit-switched technology as well as
          future ECC that are SIP-capable.

        o Emergency call mechanisms should not require a specific
          technology for determining the location of the caller.

3 Emergency URIs

   A single, global (set of) identifiers for emergency services is
   highly desirable, as it allows end system and network devices to be
   built that recognize such services and can act appropriately. Such
   actions may include restricting the functionality of the end system,
   providing special features, overriding user service constraints or
   routing session setup messages.



H. Schulzrinne                                                [Page 3]


Internet Draft                    SOS                    January 8, 2003


   UAs that determine that a dialog or transaction relates to an
   emergency MUST use an emergency call identifier in the Request-URI.
   The Request-URI MUST be either an emergency SIP URI defined in
   Section 3.1 or an emergency tel URI defined in Section 3.2.

3.1 SIP URIs for Emergency Calls

   It is RECOMMENDED that SIP-based [1] end systems and proxy servers
   support a uniform emergency call identifier, namely the reserved user
   name "sos" within any domain, e.g.,


     sip:sos@example.com
     sips:sos@example.com



   The reserved name is case-insensitive.

   The host part of the emergency URI SHOULD be the host portion of the
   address-of-record of the caller. The "sips" form SHOULD be used to
   ensure integrity and confidentiality. All SIP requests with URIs of
   this form are assumed to be emergency calls.


        The domain-of-record was chosen since a SIP user agent may
        not be able to determine the local domain it is visiting.
        This also allows each user to test this facility, as the
        user can ensure that such services are operational in his
        home domain. An outbound proxy in the visited domain can
        handle the call if it believes to be in a position to
        provide appropriate emergency services.

   In addition, we reserve user addresses beginning with the string
   "sos." for specific emergency services:


   sos.fire      fire brigade
   sos.rescue    ambulance (rescue)
   sos.marine    marine guard
   sos.police    police (law enforcement)
   sos.mountain  mountain rescue


   The sub-addresses are also case-insensitive. Additional subaddresses
   can be registered with IANA (Section 11).





H. Schulzrinne                                                [Page 4]


Internet Draft                    SOS                    January 8, 2003


        In some areas, these emergency services use different
        numbers.

   The SIP URI user name "sos" and user names starting with "sos." MUST
   NOT be assigned to any regular user.

3.2 Tel URIs for Emergency Calls

   User agents SHOULD determine the local emergency numbers, either by
   consulting their manual configuration for devices that do not move
   across national borders, by DHCP (Section 9) or some other
   configuration mechanism. If a user agent has no knowledge of local
   emergency numbers, it MUST also recognize the digit strings 000, 08,
   112, 110, 118, 119, 911 and 999 as emergency numbers.


        SIP user agents, such as Ethernet deskphones, that are
        unlikely to move frequently across national borders can
        easily implement a local dialing plan that recognizes local
        emergency numbers. Mobile devices, including PDAs and
        laptops, may not have a reliable way of determining their
        current location. Using automatic configuration avoids
        collisions with extensions that equal one of the eight
        numbers above. If a local network does not have an outbound
        proxy server, local dial plans also do not apply, so the
        problem of number collision does not arise. Collisions with
        non-emergency service numbers are still possible, albeit
        less likely. For example, 118 is used for directory
        assistance in the United Kingdom.

   If the user dials any of these digit strings, the UAC SHOULD generate
   a request with the "sos" URI described in Section 3.1 unless it has
   discovered a local outbound proxy as described in Section 9. In that
   case, a UAC MAY use a "tel" URI [3,4] without phone-context, such as

   tel:911
   tel:112



   Outbound proxy servers MUST be configurable to recognize additional
   local emergency numbers in "tel" URIs.


        There are about 60 service numbers for emergency services
        in the world; including them all is not practical, as that
        would interfere with existing local two, three and four-
        digit dialing plans.



H. Schulzrinne                                                [Page 5]


Internet Draft                    SOS                    January 8, 2003


4 Request Handling

   Once identified, a user agent SHOULD direct an emergency call request
   to an outbound proxy server or use the discovery mechanism described
   in Section 9 to find a local PSTN gateway that can connect the caller
   to a local emergency call center.

   Outbound proxy servers MUST recognize all local emergency numbers as
   well as the tel URIs enumerated in Section 3.2. The proxy MAY use any
   additional information contained in the call request, such as Mobile
   Country Code and the Mobile Network Code for 3GPP devices, to
   recognize additional numbers as emergency numbers.

   It is RECOMMENDED that gateway SIP MESSAGE requests are directed to a
   TTY-for-the-deaf translator or a short-message service (SMS) if the
   emergency call center cannot handle SIP instant messaging.


        Using a proxy server that is local to the user agent is
        more likely to reach a geographically local server,
        although that is not guaranteed if virtual private networks
        are being used.

   User agent servers and proxy servers MUST NOT require that the user
   agent client be registered or authenticated in order to place an
   emergency call.

   OPTIONS requests to the user "sos" and the "sos.*" addresses
   (sos.fire, etc.) can be used to test if the "sos" addresses are
   valid.  As in standard SIP, a 200 (OK) response indicates that the
   address was recognized and a 404 (Not found) that it was not. Such
   request cause no further action. It is RECOMMENDED that user agents
   periodically automatically check for the availability of the "sos"
   identifier and alert the user if the check fails. The period of such
   automated checks SHOULD NOT be less than once per day and MUST be
   randomly placed over the testing interval.

5 Determining User Agent Location

   Proper handling of emergency calls requires knowledge of the caller
   location to route the call to the appropriate ECC and to help the ECC
   in locating the caller when rendering emergency assistance. We
   describe the routing details in Section 6.

   The SIP UA determines its location, preferably ahead of the emergency
   call. It MAY use DHCP [9], retrieving the the location one or more
   of: geospatial coordinates (longitude, latitude and altitude) [10],
   civil address (street and community) [11] or network access



H. Schulzrinne                                                [Page 6]


Internet Draft                    SOS                    January 8, 2003


   information such as the port and switch number or the wireless cell
   identity.

   The UA needs to inform the DHCP server about its network attachment
   point. There are several possibilities, including use of the RFC 3046
   [12] Agent-Circuit-ID or Remote-ID sub-options. This approach will
   only work if the DHCP relay agent is colocated with the LAN device
   close to the SIP UA. Another option, not yet fully supported, is to
   have the UA determine the device and port information and then
   include this in the DHCP request. There currently is no DHCP option
   for doing so, however.

   The UAC inserts this location into a SIP header field. For geographic
   information, this might look something like the following:


   Location: ;lat=38.89868 ;long=-77.03723 ;alt=15 ;alt-unit=m
     ;lares=0.000122 ;lores=0.000122
     ;hno=600 ;lmk="White House" ;mcn="Washington"
     ;stn="Pennsylvania" ;sts="Ave" ;sta="DC"
     ;privacy=dnf


   Here, we assume that the DHCP option provided a resolution of 22
   bits.  The example is taken from [13].

   (The SIP header field format is fictitious and is defined in TBD.)

   For 3GPP networks, the P-Access-Network-Info header field [14] can
   convey the cell information, as defined in 3GPP TS 24.229.

   Alternatively, an outbound proxy may map the UA's device address to a
   physical location, e.g., based on a traceback within an Ethernet
   switched LAN. Such mechanisms are beyond the scope of this document.

6 Request Routing

   Any proxy, outbound or otherwise, that receives such a request MUST
   forward (proxy) or redirect the request to the appropriate emergency
   number local to the caller, using the location information described
   in Section 5.

   Note that in some limited cases, the proxy may be able to determine
   that the requestor is in the same local network even without explicit
   location information. This may be the case, for example, if the IP
   address of the request indicates a local device and the network
   offers no VPN services. Even under these restricted circumstances,
   back-to-back UAs may mislead the proxy in this estimation.



H. Schulzrinne                                                [Page 7]


Internet Draft                    SOS                    January 8, 2003


   We distinguish two cases, depending on whether the proxy has access
   to a location-to-ECC mapping service or not. A special, but
   important, case is that the caller is known to be local to a PSTN
   gateway.

6.1 Known to be Local

   In some cases, the proxy server can reliably determine that the
   caller and a local PSTN gateway are in the same emergency service
   area. In that case, the proxy forwards the call request to the
   gateway, translating the emergency URI into the local emergency
   number.

6.2 Mapping Service Available

   We refer to the location-to-ECC mapping service as a jurisdictional
   directory service (JDS) since it maps geographic and/or civil
   locations to emergency response jurisdictions. [TBD: better term?]
   The JDS can be considered a special kind of SIP location service. The
   protocol between the proxy and the JDS is beyond the scope of this
   document.


        One plausible solution simply proxies the SIP request
        itself to the JDS.

   Conceptually, the JDS is provided with a geographic location and
   possibly the type of emergency service requested (for
   "sip:sos.service" URIs) and returns SIP or tel URIs for one or more
   ECCs serving the caller. If the JDS does not recognize the service
   specification, it treats the mapping request like a general emergency
   service request.

   The tel URI returned by the JDS will contain a globally reachable
   (E.164) number, i.e., a global-number according to [4]. The proxy
   routes the call accordingly, using a local mechanism to determine the
   appropriate gateway, e.g., TRIP [5].


        Ideally, the chosen gateway should be local to the ECC, but
        that may not be achievable, as it would require a gateway
        in every community. In the United States, for example,
        there are about 5,000 primary emergency call centers,
        called Public Safety Answering Points (PSAPs).

6.3 No Mapping Service Available

   If the proxy does not have access to a JDS, it attempts to pick the



H. Schulzrinne                                                [Page 8]


Internet Draft                    SOS                    January 8, 2003


   closest PSTN gateway, translates the Request-URI to a locally valid
   emergency number and proxies the call to that gateway.

   If a proxy receives a service-specific request of the form
   "sip:sos.*@domain" (such as "sos.fire@example.com"), the proxy
   forwards it to the local appropriate specific emergency service. If
   it does not recognize the suffix (e.g., "fire"), it MUST forward the
   request to the appropriate general emergency contact, handling it as
   if the address was "sip:sos@domain".

7 Caller Identification

   When using a PSTN gateway, the gateway causes the calling number to
   be a telephone number that is mapped by the ECC to the location of
   the caller. (The process for creating such mappings is beyond the
   scope of this document. The process has been demonstrated in some
   jurisdictions for multi-line telephone systems.) It is not clear
   whether all circuit-switched trunk technologies allow potentially
   arbitrary, out-of-area calling numbers.

8 Call Behavior

   The user agent SHOULD not issue a REFER during an emergency call.

   The user agent SHOULD NOT issue a BYE request during an emergency
   call. If the user "hangs up", it is RECOMMENDED that the end system
   generate an alert tone until the user reconnects.

   The UA SHOULD automatically accept an incoming call from the same
   entity that accepted the previous emergency call.

        This allows the ECC to call back should the call be
        interrupted accidentally.

9 Identifying the Local Emergency Numbers and Gateway

   There are many ways that a user agent can configure emergency numbers
   for use in analyzing calls made with telephony-type user input. These
   include configuration tokens such as SIM cards in mobile devices, or
   protocol-based solutions. We describe one such protocol-based
   mechanism, based on DHCP, but this does not imply a requirement for
   devices.

   We propose a new DHCP option that enumerates the valid local
   emergency identifiers, as a list of "tel", "sip" or "sips" URIs.
   These identifiers can be used by the UA to trigger special behavior
   when the user dials those numbers, or they can identify a local PSTN
   gateway that can provide local emergency service. A DHCP server



H. Schulzrinne                                                [Page 9]


Internet Draft                    SOS                    January 8, 2003


   SHOULD advertise its local emergency number as well as those numbers
   among the eight digit strings enumerated in Section 3.2 that do not
   collide with local non-emergency services or extensions.

   This DHCP option MUST NOT be used if DHCP does not announce the local
   SIP server [6].


        Unlike an outbound proxy server, the DHCP server is very
        likely to be located within the same country as the user
        agent.  However, since the user agent needs to perform the
        call routing, it makes little sense to have the DHCP
        information identify a set of numbers that mean nothing
        special to the outbound proxy server. Thus, server
        identification and emergency number identification belong
        together.

   If the local network supports the location of gateways via SLP [7],
   the user agent can discover such gateways. The SLP service
   description needs to be enhanced with a list of valid emergency
   numbers.

   Details are described in TBD.

   [This is for discussion only and, if suitable, will move to a
   different draft.]

10 Alternative Identifiers Considered

   The "sos" SIP URI reserved user name proposed here follows the
   convention of RFC 2142 [15] and the "postmaster" convention
   documented in RFC 2822 [16]. One drawback is that it may conflict
   with locally assigned addresses of the form "sos@somewhere".

   There are a number of possible alternatives, each with their own set
   of advantages and problems:

        tel:sos This solution avoids name conflicts, but is not a valid
             "tel" URI. It also only works if every outbound proxy knows
             how to route requests to a proxy that can reach emergency
             services. The SIP URI proposed here only requires a user's
             home domain to be appropriately configured.

        URI parameter: One could create a special URI, such as "aor-
             domain;user=sos". This avoids the name conflict problem,
             but requires mechanism-aware user agents that are capable
             of emitting this special URI.




H. Schulzrinne                                               [Page 10]


Internet Draft                    SOS                    January 8, 2003


        Special domain: A special domain, such as "sip:fire@sos.int"
             could be used to identify emergency calls. This has similar
             properties as the "tel:sos" URI, except that it is indeed a
             valid URI.

11 IANA Considerations

   Subaddresses of the "sos" address are registered with IANA This
   specification establishes the "sos" subaddres sub-registry under
   http://www.iana.org/assignments/sip-parameters.

   Subaddresses are registered by the IANA when they are published in
   standards track RFCs. The IANA Considerations section of the RFC must
   include the following information, which appears in the IANA registry
   along with the RFC number of the publication.

        o Name of the subaddress. The name MAY be of any length, but
          SHOULD be no more than twenty characters long. The name MUST
          consist of alphanumeric characters only and is case-
          insensitive.

        o Descriptive text that describes the emergency service.

12 Security Considerations

   The SIP specification [1] details a number of security
   considerations. Security for emergency calls has conflicting goals,
   namely to make it as easy and reliable as possible to reach emergency
   services, while discouraging and possibly tracing prank calls. It
   appears unlikely that classical authentication mechanisms can be
   required by emergency call centers, but SIP proxy servers may be able
   to add identifying information.

   Given the sensitive nature of many emergency calls, it is desirable
   to use the "sips" URI to ensure transport-level confidentiality and
   integrity. However, this may cause the call to fail in some
   environments.

   Allowing the user agent to clearly and unambiguously identify
   emergency calls makes it possible for the user agent to make
   appropriate policy decisions. For example, a user agent policy may
   reveal a different amount of information to the callee when making an
   emergency call.  Local laws may affect what information network
   servers or service providers may be allowed or be required to release
   to emergency call centers. They may also base their decision on the
   user-declared destination of the call.

   Outbound proxies may need to adjust their authentication requirements



H. Schulzrinne                                               [Page 11]


Internet Draft                    SOS                    January 8, 2003


   for such emergency calls.

   It is desirable to be able to verify that the call is reaching a true
   emergency call center. The caller is unlikely to know or be able to
   obtain the public key of the destination ECC since it does not even
   know the ECC identity. The responding ECC could sign the response,
   via standard SIP S/MIME mechanisms. However, the principal in the
   certificate would appear as a random-looking domain name, such as
   admin.fayette.co.ga.us, which cannot be reliably identified as an
   ECC. Here, an attribute certificate (AC) [17] could be used to
   associate the attribute "ECC" with the SIP URI. However, it appears
   unlikely that such an Attribute Authority will emerge anytime soon,
   particularly across national borders. Alternatively, ICANN could
   create a new restricted top-level domain, such as .sos, that is open
   only to accredited emergency response entities. Clearly, this is also
   not likely in the short term.

   The caller has little choice other than to trust the outbound proxy
   server acting as an ERC or to act as its own ERC. In the former, more
   likely case, the ERC will obtain the ECC identity from a database
   source it trusts. The ERC then only has to ensure that the call
   reaches the appropriate domain, which standard TLS server
   authentication accomplishes, regardless of the domain name used by
   the ECC and without the notion of attribute certificates. Since a
   caller cannot assume that all ECCs will have valid ACs, the absence
   of such a certificate is unlikely to cause the caller to abandon the
   call in an emergency. Thus, the transitive trust model, which is
   easier to implement, appears to be a more pragmatic approach.

13 Normative References

   [1] J. Rosenberg, H. Schulzrinne, G. Camarillo, A. R. Johnston, J.
   Peterson, R. Sparks, M. Handley, and E. Schooler, "SIP: session
   initiation protocol," RFC 3261, Internet Engineering Task Force, June
   2002.

   [2] S. Bradner, "Key words for use in rfcs to indicate requirement
   levels," RFC 2119, Internet Engineering Task Force, Mar. 1997.

   [3] A. Vaha-Sipila, "Urls for telephone calls," RFC 2806, Internet
   Engineering Task Force, Apr. 2000.

   [4] H. Schulzrinne and A. Vaha-Sipila, "The tel URI for telephone
   calls," internet draft, Internet Engineering Task Force, Dec. 2002.
   Work in progress.

   [5] J. Rosenberg, H. F. Salama, and M. Squire, "Telephony routing
   over IP (TRIP)," RFC 3219, Internet Engineering Task Force, Jan.



H. Schulzrinne                                               [Page 12]


Internet Draft                    SOS                    January 8, 2003


   2002.

   [6] H. Schulzrinne, "Dynamic host configuration protocol (dhcp-for-
   ipv4) option for session initiation protocol (SIP) servers," RFC
   3361, Internet Engineering Task Force, Aug. 2002.

   [7] W. Zhao and H. Schulzrinne, "Locating ip-to-public switched
   telephone network (PSTN) telephony gateways via SLP," internet draft,
   Internet Engineering Task Force, Aug. 2002.  Work in progress.

14 Informative References

   [8] B. Campbell, J. Rosenberg, and H. Schulzrinne, eds., "Session
   initiation protocol (SIP) extension for instant messaging," RFC 3428,
   Internet Engineering Task Force, Dec. 2002.

   [9] R. E. Droms, "Dynamic host configuration protocol," RFC 2131,
   Internet Engineering Task Force, Mar. 1997.

   [10] J. Polk et al., "DHCP option for geographic location," internet
   draft, Internet Engineering Task Force, Oct. 2002.  Work in progress.

   [11] H. Schulzrinne, "DHCP option for civil location," internet
   draft, Internet Engineering Task Force, Dec. 2002.  Work in progress.

   [12] M. Patrick, "DHCP relay agent information option," RFC 3046,
   Internet Engineering Task Force, Jan. 2001.

   [13] J. Polk et al., "Semantics for DHC location object within
   GEOPRIV," internet draft, Internet Engineering Task Force, Oct.
   2002.  Work in progress.

   [14] M. Garcia, E. Henrikson, and D. L. Mills, "Private extensions
   (p-header) extensions to the session initiation protocol (SIP) for
   the 3rd-generation partnership project (3GPP)," internet draft,
   Internet Engineering Task Force, Nov. 2002.  Work in progress.

   [15] D. H. Crocker, "Mailbox names for common services, roles and
   functions," RFC 2142, Internet Engineering Task Force, May 1997.

   [16] P. Resnick, ed., "Internet message format," RFC 2822, Internet
   Engineering Task Force, Apr. 2001.

   [17] S. Farrell and R. Housley, "An Internet attribute certificate
   profile for authorization," RFC 3281, Internet Engineering Task
   Force, Apr. 2002.

15 Change History



H. Schulzrinne                                               [Page 13]


Internet Draft                    SOS                    January 8, 2003


15.1 Changes since -03

        o Added description of local discovery mechanism for finding a
          local gateway.

        o Noted that 'sos' is case-insensitive and only applies to SIP
          URIs, not other URIs.

        o Described the ECC verification options available to a caller.

        o Added local gateway discovery.

        o Added outline of how to use DHCP for configuring additional
          local emergency numbers.

        o Added 3GPP emergency numbers beyond 112 and 911.

16 Acknowledgements

   Andrew Allen, Keith Drage, Mike Pierce, James Polk, Brian Rosen and
   John Schnizlein contributed helpful comments.

17 Authors' Addresses

   Henning Schulzrinne
   Dept. of Computer Science
   Columbia University
   1214 Amsterdam Avenue
   New York, NY 10027
   USA
   electronic mail: schulzrinne@cs.columbia.edu




















H. Schulzrinne                                               [Page 14]