IP: Pac Bell says Net use may collapse phone system

[mikebw@bilow.bilow.uu.ids.net (Mike Bilow)]

  James Love wrote in a message to Mike Bilow:
  
  > When a large number of lines start having phone calls that last 
  > significantly longer than a few minutes radically shift the 
  > required ratio of lines to switching/trunking. 
  
   JL>     Maybe I'm missing something here.  Is the length of call
   JL> the ONLY  important variable?  Isn't the real issue the
   JL> number of lines that can be  used at any one time....   
   JL> like the percent of customers who are using  the telephone
   JL> at the same time?    What is this percent...for the typical 
   JL> POTS user?  jamie
  
  This is complicated, as Fred Goldstein correctly points out.  There are two
  different issues: (1) total consumption of time on the switch, which is
  correlated with number of minutes of use of each line per billing period, and
  (2) peak demand on the switch, which is correlated with number of lines.
  
  So far, as Fred said, switching capacity has been much less expensive than
  local loop capacity.  However, this depends upon the assumption that switching
  capacity need grow only logarithmically as the number of lines in order to cope
  with peak demand.  It is true that a small number of "permanent" circuits set
  up across circuit-switched service will "nail up" switching capacity at a
  faster rate than the design assumptions will allow, and this is what terrifies
  the telephone companies.  On the other hand, as I pointed out, it is the rate
  structure foisted upon us by the telephone companies themselves which
  encourages this ultimately antisocial behavior by creating strong economic
  incentives to do it.
  
  I do agree that the issue of an average four-minute call turning into an
  average five-minute call is not what the telephone companies are really
  concerned about, and that this is some sort of canard or ratemaking ruse.  The
  switching capacity demand increases only linearly with total usage, and this
  represents insignificant cost in the grand scheme of things.  On the other
  hand, as Fred also suggests, this is a good "foot in the door" to begin arguing
  for the adoption of universal measured service, the holy grail of all telephone
  companies despite its increasing irrelevance and stupidity.
  
  Regardless of its importance and real effect on plant and equipment costs, peak
  demand is essentially unrelated to average call length because peak conditions
  are only experienced for a very short time.  You can see this in cellular
  telephone rates which float to market conditions, where off-peak airtime is
  often given away free.
  
  The proper way to cope with peak demand, from a technical standpoint, is to
  charge for it explicitly.  This is how electricity has been rated for years,
  and the concept underlies several telecommnications techniques.  With frame
  relay, you are charged for a Committed Information Rate (CIR) which represents
  the minimum bandwidth guaranteed to be always available to you.  When unused
  datagram switching capacity is available, you can see substantially better
  throughput than the CIR.  This provides an economic incentive for customers to
  do the sensible thing, paying for what really costs money.
  
  As long as the telephone companies force data customers into consuming
  circuit-switched capacity instead of data-switched capacity, the rate structure
  is going to have fundamental inconsistencies and absurdities.
   
  -- Mike
  
  
  -------------------------------------------------------------------------------
  Bilow Computer Science   | +1 401 944 3937 (voice) | Michael S. Bilow
  Forty Plantations        | +1 401 944 7966 (fax)   | President
  Cranston, RI 02920-5554  | +1 401 944 8498 (BBS)   | mikebw@ids.net  (Internet)
  PGP Public Key fingerprint  =  4B 06 23 FB 3E 24 A5 24  14 B5 A2 14 96 73 B4 B2
  -------------------------------------------------------------------------------