North American Network Operators Group

Date Prev | Date Next | Date Index | Thread Index | Author Index | Historical

Re: RWHOIS, SWIP, and proving ownership

  • From: David R. Conrad
  • Date: Thu Nov 21 01:32:43 1996

[note reply-to line]


I'm curious -- given the registries are required to conserve address
space, what would you suggest the registries do to attempt to insure
address space allocated is actually used?

>Consider this a side note to the preceding discussion.
>1. Most of our clients understand that their ''lease'' on network address
>   space is at our whim, by contract for IP connectivity, and is subject
>   to renumbering if WE assign them new space, if WE are assigned new space,
>   or if they move elsewhere.  
>   Therefore they don't ask us to SWIP the nets nor do they care.  
>2. RWHOIS doesn't run on any production operating system.  I know Unix is
>   in vogue, but since we do the 99.96% uptime schtick, we use operating
>   systems that stay up (VMS).  This means we can't run RWHOIS (even if we
>   did want to, which if you read #1 above you'll see we don't.)
>3. We currently use almost all of a /18, two thirds of a /19, a few /22s,
>   and some /24s.  It would be easy to justify a /17 based on all this, but
>   if someone wanted to be rigid about RWHOIS and SWIP, even a bunch of 
>   traceroutes aren't going to convince them.
>Back in THE GOOD OLD DAYS (tm), we said "Be flexible with what you accept,
>be rigid with what you send out."  (Others made it sound better and put it
>in RFCs... D.C. for one.. :)
>Nowadays I see the motto has become "Be rigid in what you accept, and modify
>your templates as often as possible."  This criticism applies equally to the
>RA IRR as it does to the InterNIC.
>Gee, and this started out as one sentence that went "We don't run RWHOIS, our
>clients don't want it, our operating system won't support it, and you better
>listen when we ask for a /17 ;)"
- - - - - - - - - - - - - - - - -