Netmask question

[Barry Margolin]

In article <Pine.LNX.3.96.1000227172511.17284A-100000 at brian.vpszk.bme.hu>,
Ferenc Tamas Gyurcsan  <gyurex at brian.vpszk.bme.hu> wrote:
>Hi,
>A question came up, and I can't find the answer neither in the rfc's, nor

Why are you posting this in the BIND newsgroup rather than the TCP-IP
newsgroup?  I have redirected followups to comp.protocols.tcp-ip.

>in the faq. So is it possible to have a netmask that's not contigous? I
>mean that in the end for example something like this:
>.11000011
>So it would describe 255.255.255.195. Where can I find info about this? An
>rtfm answer would be great, showing which fm. Somewhere I read it was
>possible, and many people said it wasn't.

It used to be possible, but these days it generally isn't.  The general
scheme of network masks has been superceded by CIDR, which uses prefix
lengths.  A prefix length is just a value from 0 to 32, specifying the
number of high-order bits in the address that represent the network; it
corresponds to the number of contiguous 1 bits in the network mask.  Most
systems still require you to enter masks, but they translate them
internally into lengths.

This simpler system is required for determining which prefixes are more or
less specific than others.

-- 
Barry Margolin, barmar at bbnplanet.com
GTE Internetworking, Powered by BBN, Burlington, MA
*** DON'T SEND TECHNICAL QUESTIONS DIRECTLY TO ME, post them to newsgroups.
Please DON'T copy followups to me -- I'll assume it wasn't posted to the group.