RE: types of traffic in tcp congest control

["Douglas Otis" <[e-mail]>]

Anumita,

Core aspects may deserve separate consideration:
 Long distance
  - Transoceanic
  - Transcontinental
  - Satellite
 Metropolitan
  - Central Office
  - Head-end
  - Pax
  - Cable

Provisioning of these aspects create the Core and each have respective
limitations based on deployment.  Last mile technologies have seen
reductions in basic service to assist in with Core issues as improved access
(last mile) technologies become popular.  Although technology allows
phenomenal bandwidth, this is not without a requisite investment.  Early
investments in this technology were stranded as growth slowed.  Expectations
of improvement in basic service have diminished as have investments.

A good Core network must be over-provisioned to allow fail-over and,
although routers are better at handling saturation, lack of free bandwidth
does have a dramatic service impact.  There is a significant difference in
the level of basic service between countries owning to adequate essential
Core and population densities that constrain cost.  Some aspect of the Core
is a constraint where small changes in access will absorb available
bandwidth.  Web caching is an effort to relieve Core constraints, as
example.

Will greatly improved CPU and LAN technologies press for improved basic
access?  What model for Core provisioning is used to keep pace?  To use the
network to enjoy multimedia, then the present 1 mb/s must be improved to a
reliable 6 mb/s for broadcast quality, and 11 to 22 mb/s for high
definition.

The many Mice (low-rate last mile) will still scare away the Elephants if
sharing under provisioned Core.

 -Doug

On May 14, 2002 11:31 PM Anumita Biswas ([e-mail]) wrote:
>
> Can the arguments about disk drives, memory capacity be really extended
> to core, highly shared Internet links? Has the provisioning of core
> links ever exceeded the rate at which the number of internet users are
> increasing.
>
> Also, should we not also consider the "bandwidth" of the core
> routers/switches, that is the rate at which their ability to
> route/switch packets is increasing versus the rate at which they are
> pumped with packets? Can that ability be overprovisioned quickly
> enough?
>
> Please revert back, if I have misunderstood.
>
> thanks
> Anumita.
> >>> Joe Touch <[e-mail]> 05/15/02 11:18AM >>>
>
>
> Anumita Biswas wrote:
> > Is it not true that overprovisioning is never a long term solution,
> > as bandwidth is never ever enough? As Alan says, "Usage expands to
> > fit network bandwidth".
>
> If you have a DSL line, track its usage. You will find that ~1Mbps is
> more than sufficient most of the time.
>
> Yes, applications sometimes catch up. However, whenever a resource is
> scarce the assumption is that usage will always catch up. The arguments
> held in the past for disk drives, memory capacity, and CPU speed. Most
> are, for most purposes, overprovisioned very easily for all but a small
> fraction of the time for most users.
>
> Joe
>
>
>