!! (fwd)

Wed Feb 2 16:06:04 UTC 1994

Richard,

  I agree that we must be careful.  However, we must setup guidlines that 
  have a technical basis. If we are to achive efficent CIDRization of the 
  address space then we can not give  CIDR  blocks to anyone who says 
  they are a provider. I don't think that we have ever denined anyone 
  address space. The complaint from Kansas was a person that requested 
  space for Kansas's network and was given a block of 512 through MIDNET. 
  MIDNET was the provider at the higher level. Now he wants to be a provider 
  for himself connected to ???.  I understand that the look of what is 
  currently the NSFnet will change. What if tommorow I want to become a 
  provider and request MY block of 256.  I think we must ask the questions 
  and make a smart allocation. Souldn't the provider to which he will
  connect be involved in the allocation? I understand that anyone should 
  be able to switch providers at any level. I would assume that the block 
  could go with them.

Scott

> 
> Scott,
> 
> > I have asked the question many time.  "What is a provider?".  Once the 
> > CIDR allocation started the "Providers" came out of the woodworks.
> > No one so far has given an answer to the question that the majority can 
> > agree with. I will not be at the regional tech meeting in CA but Mark will. 
> > I don't know a group better suited to answer the question that established
> > providers. Do us a favor and come up with a proposed answer to "What is a 
> > provider".  I will work with NSF and Postel to make it policy. This would
> > make our life easier.
> 
> I think we have to be very careful here.  When we make policy that
> adversely affects someone's business interests we're just begging for
> a suit.  Especially with the *perceived* "shortage" or IPv4
> addresses.  It seems to me that it would be very difficult to exclude
> anybody from the category of "provider"  (for the purpose of giving
> him or her addresses) who has even the flimsiest claim to being one.
> Perhaps some sliding scale based on demonstrated need would work.
> But it would have to be conservative (liberal?) and easily quantified
> to make it defensible.
> 
> --Richard
> 