Re: VCC cost models .... (was Re: Limits

["Silverman, Steve" <[e-mail]>]

There is apparently a widespread assumption that doing an SVC "shortcut or 
cut-thru
will create an additional cost.   If this is true, I doubt that any of this
will get deployed.  I think most of the IP over ATM enthusiasts believe it 
will
cut rather than increase costs.

I think there is confusion over price vs. costs.   The Internet has always
been cheap in price because the US government subsidized it.  If you knew 
what
it actually cost during the early '80s,  those of you who payed those taxes 
would howl!

On the other hand, the U.S. has a long tradition of having long distance
subsidize access costs via a high usage charge far beyond the cost of 
service delivery.
This was a deliberate decision by govt. regulators which was undercut over 
the last 15 years by
the creation of competition among IXCs but the philosophy lives on in 
certain ways.

The assumption is that call setup is a large cost in creating a VC.

For an IP packet: the first packet involves a routing decision and caching 
that decision for each switch.

For a VC, the first packet involves a routing decision (considering traffic 
load in the net), caching that decision
in each switch, and setting up the call accounting info so the user can be 
charged.

In order to do decent QoS, the IP network will have to start considering 
traffic load.
If the VC network did not do usage billing, and simply did no call 
accounting, the cost associated with
the call setup is not that different than the cost of an IP "session". 
 (Look out for TELCO people
committing suicide at the mention of eliminating the usage charge. :-)   ) 

We will need a fair amount of computer power to do call control but that is 
only
a "Small Matter of Silicon" and so far the price of silicon has dropped 
almost as fast as the
software gobbled it up.  Call control would appear to lend itself to 
parallel processing.

In my opinion, one reason for the growth of certain services in the U.S. is 
flat rate charging.
If this industry is going to continue to grow, one helpful factor would 
appear to be such flat
rate charging.  If every Web page click is answered with a reminder of an 
additional charge,
usage will be low, and we are wasting our time.


My personal opinion, not my employer's
Steve Silverman