TELCAN TOLL FREE
Real-time Activation.
Change ring-to number in real-time.
Real Time Downloadable Call Detail.
Ring-To Numbers Anywhere in the World.
Online Account Management.

ENUM - THE OTHER POINT OF VIEW

John C. Klensin is an Internet founder who helped design the first email and file-exchange systems.

by John C Klensin, presented at the APT-ITU Joint Workshop on ENUM and IDN in Bangkok on August 25, 2003

Participants should, in designing policies about ENUM, be aware that there are two points of view about it that have not been well-represented in any of the efforts discussed here. Both the ITU-T and IETF efforts have involved work by ENUM proponents to define a good way to do it given that it is to be done. Neither IETF nor ITU-T have taken a formal position that ENUM (or any other E.164-based scheme for "Internet telephony") is a good idea and should be implemented by member states.

These notes are not intended to argue for these alternate points of view, but to make attendees aware that they exist. The negative positions can be summarized as:

(1) ENUM is a system for accessing PSTN resources from the Internet and possibly Internet resources from the Internet. It is needed there precisely because Internet applications must use the domain name system, directly or indirectly, to resolve names (or telephone numbers) into addresses of systems that will handle the relevant traffic. For ENUM, the DNS must be used to identify another Internet host or the Internet side of a gateway to the PSTN and may, with the latest proposed version, be used to identify particular services being offered at particular numbers.

ENUM can actually be avoided in this case, by having the user-level protocol explicitly identify the gateway and routing to be used. However, the long-term experience on both the PSTN and the Internet is that forcing users to identify particular routes or particular gateways causes serious user troubles and even more severe difficulties in designing, maintaining, and operating a secure, reliable, and robust network.

From the PSTN origination side, ENUM is not needed and may be too complicated. If used inappropriately, it may even cause violations of existing recommendations, service definitions, and regulations about, e.g., maximum call connect times. A number dialed in the PSTN must pass through a switch with appropriate routing capability. That switch must be able to "figure out" how to handle the number -- forwarding to another switch, termination on an instrument or PSTN terminal ("telephone"), or passing to a gateway to another environment. While it is possible in theory for a switch to make this determination by looking up all numbers in the DNS, assuming that they are to be terminated within the PSTN if no DNS record is found, that approach would be pathological in practice: not only will there be more PSTN-terminated numbers than Internet-terminated ones for the foreseeable future, but DNS timeout cycles (often required to determine that a number does not exist as an ENUM record) are in multiples of seconds -- far too long to wait to determine that a number should actually be terminated on the PSTN.

Suppose, instead, that the switch can somehow be used to determine whether a particular number is to be routed to the PSTN or to a telephony - Internet gateway (needed if only to do protocol conversion). While it is possible for such a gateway to do an ENUM DNS lookup to map the number into an appropriate Internet-terminated URI, it is equally possible for the gateway (or the associated switch) to maintain its own mapping table from the E.164 number to an Internet resource. Maintaining such a mapping in the switch is preferable for a number of reasons, including its being more responsive to changes within the PSTN (number portability, call forwarding, reallocation of numbers) than difficult coordination of PSTN numbering/routing systems with DNS entries.

Of course, if one allocates a particular number block to Internet telephony (e.g., with a "city code") , the PSTN switch would know that all such numbers were to be terminated on the Internet without explicit routing to an Internet telephony gateway. But this does not really change the principles above. More important, it would quickly violate principles of number portability in countries that consider such portability important, since it would be impossible to migrate between PSDN and Internet service without changing numbers.

So, while ENUM can be used in routing from the PSTN to the Internet, it is not necessary and may not be desirable.

(2) As telephony and telephony-like services continue to advance, it will become increasingly attractive to identify a person or function to be reached, not to continue to simulate a copper pair terminating on a shared instrument. One rarely wants to reach such an instrument or a simulator for it, one wants to reach a person. More important, that desire to reach a person is typically tied to access priorities, and not to selecting from among a list of numbers for business or home telephones, numbers for an assistant, fax numbers, mobile numbers, and so on. Some of the machinery that underlies ENUM -- particularly DNS NAPTR records with service-type identifiers-- may be very useful in supporting such advanced services. They may be even more useful if used in conjunction with a high-capability service requesting device, like an alphanumeric Internet terminal, rather than with a telephone keypad. But, in this case, the E.164-like number is less desirable than using the proposed recipient's name and other identifying information to access an appropriate directory database.

Copyright © 1995-2009 ICB. Inc. All rights reserved. "ICB Toll Free News" is a trademark of ICB Inc. ICB Inc. assumes no responsibility for use or misuse of information contained herein and / or accessible via this site. ICB Inc. cannot and does not vouch for the accuracy and / or usability of any of the contained or linked-to information, and assumes no responsibility for any errors or omissions.