On Sun, 17 Feb 2002, Paul Courchene wrote:
> "[Bruce Mohl is] seeking volunteers with high-speed Internet access with
> AT&T, RCN, Verizon, and other DSL providers who would be willing to record
> their actual Internet "throughput" and compare it to the speeds promised."

  Heh.

  You have to realize something about Internet feeds.  The bandwidth of the
"last hop" link is but one of many factors involved.  To illustrate with an
extreme example, a T1 feed won't do you much good if the router it is
connected to is in turn fed by a 2400 baud modem.

  ISPs over-subscribe their networks.  Take, for example, a 144 kilobit/sec
DSL.  That DSL is between your house and the ISP's equipment in the local
CO.  The ISP will have any number of subscribers also being serviced at that
CO.  All those subscribers will share the bandwidth available at the CO.  
If the ISP is feeding the CO with a single T1, all it takes is eleven people
with 144 kilobit/sec links before the potential subscriber load exceeds the
bandwidth available at the CO.  (T1 = 1536 kilobit/second.  11 * 144 = 1584.  
1584 > 1536.)

  This situation repeats itself at each point of concentration within the
ISP's network.  Eventually, you reach the ISP's own provider(s), and the
situation continues to repeat, except now your ISP is in your shoes.  Once
you reach large ISPs, you instead find peering points which connect their
networks to other large ISPs.  Eventually, you start following the route
"down" to your destination.

  You also have to count protocol overhead.  Some amount of your feed's
bandwidth is used to deliver you an IP transport.  Some of *that* is eaten
up by the user transport (TCP or UDP).  The application itself (HTTP, FTP,
whatever) likely has its own overhead.

  And who says the other end even has the capability to feed you at the
maximum theoretical rate your feed allows?

  So, you see, things are a good deal more complex than the theoretical data
rate of your 144 kilobit/sec DSL feed.  The only consistent measurement you
can make is from your house to the equipment at the other end of that DSL,
before your load is aggregated with other subscribers.  Getting a good
measurement of that will likely require working directly with your DSL
provider.

  Now take a look at the service agreement you got when you signed up with
your DSL provider.  I imagine it likely disclaims any and all performance
guarantees what-so-ever.  Most people sign-up for a completely un-guaranteed
service, and get exactly that.

  Go to your ISP and say you want a service guarantee, and they will
probably be happy to provide one -- at five times the price.

  I am afraid it comes down to "Good, fast, cheap: Pick any two."  When you
sign up for a cheap, fast Internet feed, you get exactly that.

  The *REAL* question you (or Bruce Mohl) should be asking is: Why do good,
fast feeds cost so much?

  Well, there are a lot of reasons, but most of the cost can be traced to
local line charges.  Verizon charges three arms and two legs for 'em.  
Having a private corporation control an inherently single-source commodity
means they can set the price.

  When you consider that Verizon is *also* in the business of selling data
services over those local lines -- which their competition must buy from
Verizon -- the conflict of interest becomes obvious.

-- 
Ben Scott <[EMAIL PROTECTED]>
| The opinions expressed in this message are those of the author and do not |
| necessarily represent the views or policy of any other person, entity or  |
| organization.  All information is provided without warranty of any kind.  |





*****************************************************************
To unsubscribe from this list, send mail to [EMAIL PROTECTED]
with the text 'unsubscribe gnhlug' in the message body.
*****************************************************************

Reply via email to