At 11:04 AM 4/4/96, Tim Bass (PIER) wrote:
Howard states:
[ ... ]
IMHO, I don't think you can guarantee that almost anything will stay routable, certainly anything less than an /18. Bluntly, there's no good way to guarantee routable prefixes.
Yes, there is a good way to GUARANTEE routable prefixes. Intermediate system address translation which maps to "virtual renumbering"....
In the specific case, this might take a new translator box that is not budgeted. Yes, I know renumbering costs staff time, but many organizations find it much easier to justify staff time than capital expenses. There's already a commitment to renumber here. Tim, my major point was that the enterprise here should, when renumbering, install infrastructure that makes possible future renumbering easier. Are you opposed to this? DHCP and such exist today, but I'm not sure I would say fully functional and tested address translators do. They may very well be in the near term. I'll agree that address translation will work in some, but not all situations. The problems are less in pure routing than in network management, DNS (especially reverse mapping), etc.
The answer my friend.... is blowing in the wind... in the technical implementation, not "human attitude re-engineering".....
I don't think we are in philosophical disagreement, but what the technical implementation should or can be.
The attitude, .... "bluntly, there's no good way for man to fly...." is not unique to written history.... it is alive and well in the Internet today ;-)
There is no good way for man to fly. There is an excellent way for man to be flown in airplanes. There isn't always a need to fly. Requirements analysis can fix things at the end system, with application-dependent solutions ranging from videoconferencing to ICBMs.
Howard, Quite frankly, I am not opposed to *anything* very much, EXCEPT for blanket statements that clearly state that the *only* solution is end-user-renumbering. NAT boxes, are not 'my-cup-o'-tea' as well, but it does offer a new commercial market for hardware vendors, and does help bandaid some problem areas. The *better* approach, IMO, is router configuration parameters in that do the translation: ip translation .... ..... ^Z write mem write net ..... No matter and irregardless of the exact implementation, there are numerous opportunites for aggregating IP routes in routers.... hopefully it will not become a marketing scheme to sell everyone NATs ..... and hence the mantra.... CIDR, renumber, NAT or DIE..... (please, not again...!!) It is not rocket science to do address translation in YFRM routers..... but there is not a lot of money in it. The hooks to make it work in the Big I is not impossible as well....... My only point is.... kindly stated..... there are more ways than one, in the Southern venacular, to 'skin a cat'.... See ya'll later :-) Tim
participants (2)
-
hcb@clark.net
-
Tim Bass