Wallace et all, an article from Bob Metcalf about IP telephony.
June 15, 1998
Is the Internet-traffic bubble about to burst? Some say telephony or bust
When, if ever, will the Internet's traffic bubble burst?
The number of Internet users will grow by another 100 million this year. Ignoring the sad fact that after 122 years most of us 5 billion humans still don't have telephones, adding Internet users could drag on for decades.
The time users spend online is going up, most of it coming out of the time people spend in front of the television. Considering how much TV we watch, there's no end in sight to increasing Internet time.
And the Internet has been taking over one huge bandwidth-consuming application after another. First came e-mail. Now the Internet carries more messages than the U.S. Postal Service carries letters. Then came Web publishing and commerce, which have been ramping up big time since 1993. And, with $200 billion a year up for grabs, the Internet, beginning with international fax, is taking over telephony. Or is it?
In my June 1 column (see "Sparks fly around Internet-telephone convergence at Vortex98"), I argued that Internet-telephone convergence is this year's biggest story in information technology. I reported on the Vortex98 conference (http://www.vortex98.com), which was convened to investigate this trend.
In my Vortex98 report I opined that the Internet will soon carry telephone calls well enough to compete with, if not immediately obsolete, the telephone network. This week I'd like to worry out loud whether this is actually true. Can the Internet carry telephone traffic with the quality and scale required to capture a good part of the telephone industry's $200 billion?
Experts on Vortex98's quality of service (QOS) panel were unanimous in the affirmative, but I'd like them to get back to me on the following doubts:
Regulation. The telephone industry is asking that Internet telephony be prohibited, and if not prohibited, then regulated to death, and when finally regulated, then heavily taxed. Too bad that just when we need their clout in Washington, Microsoft and Intel are making themselves unpopular there.
Standardization. Internet telephony needs standards. Sometimes it seems that we have plenty: T.120, H.320, H.323, H.324, RTP, RTCP, RSVP, and IPv6. The International Telecommunication Union (http://www.itu.org) is active. The Internet Engineering Task Force (http://www.ietf.org) is cranking up. Though Intel and Microsoft are singing a standards tune for the moment in the Voice Over Internet Protocol Forum (http://www.imtc.org), they will grow impatient when the IETF and ITU return to fighting. Intel and Microsoft have histories of exploiting their own standards instead of truly open ones.
Billing. Much like telephone calls, Internet calls will have to be set up, taken down, and billed out. Confusion reigns in the Internet and telephone industries over billing. Telephone billing is going flat. Internet billing is going metered. Telephone companies have settlements systems. Internet service providers don't. Billing has to be worked out, but who is doing that?
QOS. There are major problems with getting telephone QOS through the Internet. Nobody is close to solving the problem of maintaining QOS among ISPs. Peering agreements among ISPs don't even have a settlements system, let alone a system for guaranteeing one another's committed QOS. And even within one ISP, QOS is not guaranteed. When I ask ISPs about this, they say that their current plan is to throw bandwidth at the problem.
So this becomes my first Internet-telephone convergence scenario, that the Internet will capture telephone traffic without robust QOS, simply with excess bandwidth.
Internet QOS is badly broken. True, moving from "best efforts" to traffic prioritization will be progress when deployed, but we really can't afford to lose telephone service during the Internet's busy hours when there's more highest-priority traffic than capacity.
What's needed is QOS based on reservations of capacity, not just prioritizations of traffic. Deployment of this is nowhere in sight. Please let me know if you have some good news to the contrary.
My second Internet-telephone convergence scenario has the Internet continuing to burgeon, but without telephony. Publishing, commerce, and fax may be enough for a long time to come.
My third scenario, should the Internet fail to carry telephone, is that the Internet's traffic bubble will burst.
Scott Bradner, Harvard's Internet guru at the IETF, says carrying telephone would only double the Internet's current traffic, and that is already happening every four months, so what's the big deal?
Hiram
|
