>Hi groupies !!!
> I just happened upon this' whilst trying to rest
>my brain from BGP.
> A RIP II update sends a network mask out onto the
>wire for every network the router is advertising.
>
> Why didn't the designers just code the network
>mask as a 6 bit field which could represent any mask
>and save three bytes per network route ?
I can't speak exactly for the designers, but several guesses.
The original RIP I specifications allocate a full 32 bits.
While having a 6 bit field indeed would save memory/bandwidth,
it would be more processor intensive. To get the prefix from
a 32-bit mask, all you need to do is AND the address with the
mask. On complex instruction set machines, this is a single
instruction, and on RISC machines, it still is cheaper than
the looping that a "counter" style mask would need.
>
> I've had a quick fiddle trying to create some
>bizzar masks e.g 1.1.1.0 255.0.255.0 but they all
>generate errors. 'Bad Mask etc'
>
Noncontiguous subnet masks break CIDR, and are forbidden
by RFC 1812. Some host implementations will accept them,
but there's really no useful application for them.
_________________________________
FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html
Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
