[LINK] World's biggest luddite strikes again! / Content is ha
Mon, 29 Oct 2001 10:12:39 +1100 (EST)
On Mon, 29 Oct 2001, Chirgwin, Richard wrote:
> 1) The effectiveness of widespread broadband will last only until Website
> designers find ways to make a 6M link act like a 2400bps modem.
sure :) In general, increasing the amount of bandwidth available results
in more bandwidth utilization. No argument there
> 2) Saliya, nobody, but nobody, gets 100% utilisation out of any link.
> Nobody, never. Try this experiment: take a broadband link, work out the
Of course they do (well, say to within 1-2% or so). Even though we weren't
talking about 100% utilization of a broadband link, and even though it's
far easier to approach 100% on a 56k modem, the principle still applies.
This will happen when the remote site(s) and the pipes between them/you
have sufficient capacity and allocation to flood *your* hop.
A single download from a single site may not flood your link; increasing
the number of remote sites accessed and the number of simultaneous
downloads increases the chances of this happening. You can do this
manually, or it can happen automatically if you use some type of download
Increasing the amount of available bandwidth simply means that the
*chances* of 100% utilization happening for any given flow are lower (but
> theoretical download time of, say, a txt file from project Gutenberg; then
> run the download with a stopwatch. I'll bet anything up to five dollars that
> the actual download time is rarely better than around 40% theoretical.
> (which is another point that probably holods consumers back: carriers
> advertising X megabits per second, but delivering X/50 actual).
Most ISPs (I own one) buy bandwidth at X Mb/s and (here is the important
part) *monitor* this utilization. If we were paying for 45Mbps but only
getting 1, we would be demanding to know why.
What you are probably talking about is oversubscription; where the
outbound link capacity is exceeded by the sum of the inbound links.
If your ISP does this, and it affects you, find another one. Most ISPs
oversubscribe their links; the better ones will use link utilization to
tell them when saturation is reached.
> 3) Consumers don't apply $X/hr to recreational activity.
Sure, but I'd be surprised to find *anyone* whose use of the Internet was