[e2e] Re: crippled Internet

[Christian Huitema]

The business model of providing poorer quality for a lower price is not
very attractive, especially when the "higher price" is a few cents per
minute. Ike Elliott of Level3 describes it as "less for less" -- quite a
good summary. Today, we have dual networking -- one infrastructure for
voice, another one for IP. There is no question that if we want to move
to a single infrastructure, we must demonstrate that IP communication
can be both better and cheaper than traditional telephony -- "more for
less." 

There are in fact many reasons why IP communication can surpass
traditional telephony. First, going from TDM to packet makes it easy to
provide better sound quality -- move from 8KHz sampling to 16KHz
sampling, or more; in fact, using 16KHz sampling with a G.722 codec
requires less bandwidth than traditional TDM call. Second, IP
communication can easily accommodate multiple media at the same time,
such as audio, video, application sharing. Maybe even more important, IP
communication enables us to immerse telephony in a general data
experience, tying the processing of calls with instant messaging systems
or with application databases.

Good quality requires adequate bandwidth and short delays. In practice,
the bandwidth constraint is well met in "modern" networks; it is not met
today in some under-provisioned countries, but this is arguably a matter
of build-up and time. So, if we could meet the delay constraint, life
would be good. As Craig mention, this implies "mouth-to-hear" delays of
less than 150 for acceptable quality -- higher delays have been shown to
impact the flow of the conversation. The mouth to hear delay includes
processing, such as echo cancellation and compression, packetization,
network transmission and play-out buffers to compensate the jitter.

Looking at the NLANR data (http://watt.nlanr.net/active/) we see
measurements such as "average RTT of 60 ms plus standard deviation of 5
ms", one may be tempted to conclude that life is good -- average one way
delay plus two std sums to 40 ms, much less than 150. However, there are
two caveats. First, a significant portion of the 150 ms budget must be
reserved for the processing function -- maybe half of it in the current
state of the art; this leaves a target of 75 ms for the network delay.
Second, measuring the std is a very poor approximation of the behavior
of jitter compensation algorithms. We need a much finer grain
understanding of the structure of the delay.

Stanislav mentions that "Surveyor shows 0th, 50th, and 90th percentile
graphs by default." It is a good start, but there are two issues. First,
surveyor uses the IPPM defined methodology, sending probes at
pseudo-random intervals; this is not a good modeling of the VoIP
traffic, which mostly consists of "talk spurts", trains of packets sent
at a 20 ms or 30 ms interval; the interesting measurement is the maximum
cumulative jitter within a packet train. Second, even if we accepted
random sampling, we would have to check for the 99th percentile, rather
than the 90th percentile. We could only derive this percentile from the
surveyor data if we new the structure of the delay distribution: if the
distribution is Poisson, then the 99th percentile will be about 2.4 time
farther from the mean than the 90th percentile; if it is a power law
with a 1.5 exponent, it will be 6.5 times farther.

Can we get better data?


> -----Original Message-----
> From: Sean Doran [mailto:smd at ebone.net]
> Sent: Thursday, April 26, 2001 7:41 AM
> To: J.Crowcroft at cs.ucl.ac.uk; vjs at calcite.rhyolite.com
> Cc: end2end-interest at postel.org
> Subject: Re: [e2e] Re: crippled Internet
> 
> So is it fair to ask if there is a relationship between people's
> acceptance of poorer quality when the charging regime is very
different
> (and cheaper), just as there is an apparent relationship between
> people's acceptance of poorer quality when there is a substantial
> feature change (e.g., GSM in general)?
> 
> It's not like I'm just saying, "let the market decide", but rather
> wondering if anyone has good guesses at what the decision might be
> if ISPs focused on offering good-enough-for-data quality as usual,
> rather than tightening things up to some ITU standard for
delay&jitter.
> 
> 	Sean.