RE: Newbie Question about addressing impacts

Prefix conservation or any type of forced VPN within the answer with NAT
is immoral view as technologist.
/jim 

> -----Original Message-----
> From: [email protected] 
> [mailto:[email protected]] On Behalf Of Tony Li
> Sent: Sunday, August 15, 2004 4:19 AM
> To: Christian Huitema
> Cc: Fleischman, Eric; Brian E Carpenter; Multi6
> Subject: Re: Newbie Question about addressing impacts
> 
> > I am sorry, but I don't get the VPN point. The number of VPN is 
> > proportional to the number of sites willing to establish VPN 
> > connections, and not linked to the availability of global addresses 
> > for these sites.
> 
> 
> If PI addresses are made widely available, then the routing 
> table will grow.  As the routing table grows, each carrier 
> must grow its tables to support default-free routing.  To be 
> able to provide VPN services, the carrier must be able to 
> support all of the routes that his customers wish to support. 
>  Since several of these customers could reasonably be Tier 1 
> backbones, they will be carrying the full table.  Thus, there 
> is an amplification factor (currently about 10x) of the size 
> of the table vs. what routers have to hold.  If that grows 
> (as seems likely as the size of the DFZ grows), then the 
> amplification times the natural growth of the table will 
> continue to push technology.
> 
> >
> > Eric is making a basic point of economics. Large companies 
> will want 
> > to get a global address, effectively shifting network 
> management cost 
> > from their IT department to the ISP community. A number of folks in 
> > this group believe this is a bad idea, and are essentially 
> trying to 
> > write in the standards that such cost shifting should be illegal. I 
> > personally don't think it should be illegal, although I am ready to 
> > admit that it should be expensive, as in "no free lunch".
> 
> 
> Well, I don't think that "illegal" is possible, given that we 
> are not a legislative entity.  I don't think that anyone 
> would object to a market for routing table entries, but it 
> has yet to appear.
> 
> >
> > Your mention that routers commonly support 1E6 routing 
> entries seems 
> > to indicate that some amount of "cost shifting" is indeed feasible.
> 
> 
> No, I don't think that's sufficient.  We have yet to see any 
> type of cost shifting occur.  For that to happen, we would 
> have to have a market and the fact that it does NOT exist 
> means that we don't have a mechanism to work with.  Yes, this 
> is an economic, not technical issue.  However, we as 
> technologists have the responsibility to raise the awareness 
> of this and to promote prefix conservation.
> 
> 
> Tony
> 
> 
>