X-Mozilla-Status: 0001
        X-Mozilla-Status2: 00000000
               Message-ID: <3622ABB9.2F56B36A@concentric.net>
                     Date: Mon, 12 Oct 1998 21:24:09 -0400
                     From: Eric Durant <eadurant@concentric.net>
                 X-Mailer: Mozilla 4.5b2 [en] (WinNT; I)
        X-Accept-Language: en
             MIME-Version: 1.0
               Newsgroups: execpc.general
                  Subject: Re: ISDN Question
               References: <6vj4b3$ncj@newsops.execpc.com>
                           <6vja7v$982@newsops.execpc.com>
                           <6vjnd3$mmo@newsops.execpc.com>
                           <6vlqju$d5v@newsops.execpc.com>
                           <6vu0is$ied@newsops.execpc.com>
             Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

> > Robert Koch (nemesis@execpc.com) wrote:
> > : Latency is greatly affected by ISDN, as I can get 40-50ms latency to places
> > : as far away as new york, if the routing is good

> In article <6vlqju$d5v@newsops.execpc.com>, madings@earth.execpc.com
> says...
> > Really?  How is that possible?  Latency should have more to do with
> > the speed the signal travels at, which is the speed of electricity
> > (light) regardless of whether that signal is analog or digital.  The
> > extra time delay above and beyond that will be related to factors that
> > have nothing to do with the bandwith of the transmission - lag at the
> > transfer points, and so on.  Or so I thought.  Maybe what you are
> > seeing is the overhead bandwith time in the packet wrappers around the
> > data, rather than the bandwith time of the data itself.  I'm not sure
> > if that is defined as part of the "latency" time or not.

Chris wrote:
> Latency is determined by the distance (186,000 miles per second) and the
> type of connection you use. Analog-digital conversions add latency which
> all Quake players are painfully aware.

That's true, if somewhat misleading.

186,000 miles/s is the speed of an electromagnetic wave in a vacuum and
is how fast (e.g.) a radio or satellite signal propagates.  But,
propagation of wire bound signals is much slower.  Without straying even
further off topic into the relevant physics, electromagnetic waves
propagate a bit like water out of a high-pressure hose and wire bound
signals propagate like waves in a pond.  You can't take that analogy too
far, but it makes the point.

A/D and D/A conversions do increase (and may be the primary component
of) latency (due to physical factors and coding/processing) - this
happens at each switch/router/etc. on the data network, and at various
stages between the user and the telco.  I'm not an expert on ISDN, but I
believe one of the major reasons ISDN data has lower latency in general
than modem data is that the telco handles ISDN data as a data stream on
its network, while it usually converts a analog (POTS) signal to a data
stream (analog coded 1s and 0s in some frequency and/or temporal region
of the transport medium's bandwidth) for transport on its network, and
back to an audio band analog signal on the "last mile" to the call
destination.  Also, the terminal adapters are coding and decoding the
data, not modulating and demodulating it, further reducing latency.

Eric Durant
---
Website:  http://www.edurant.com/
ICQ:      9118339