From: Anthony Pace (anthonypace@xxxxxxxxxxx)
Date: Sun Aug 11 2002 - 19:59:56 GMT-3
Can anyone list an example of an "all subnets" broadcast? local
broadcast and directed broadcast, I understand. When would
10.255.255.255 have meant a directed broadcast to all subnets of the 10
network?
Anthony Pace
On Fri, 9 Aug 2002 15:02:01 -0500, "Wade Edwards"
<wade.edwards@powerupnetworks.com> said:
> Unless this is a really really really old router then this is not
> actually a bad practice anymore because it has to deal with Classful
> routing and has no bearing upon modern, within the last decade,
> networking.
>
> I assume that since sending a packet to 10.255.255.255, on a modern
> router, will only send a directed broadcast to the highest subnet on
> the
> 10-net, or be dropped as it should because of DDoS, then this is still
> not a problem.
>
> Does anyone know of any network equipment that still routes based on
> classful networks, in production, with which this would be a problem?
>
> The only reason why I bring this up is because some people that were
> studying for their MCSE brought up to me that the top and bottom
> subnet,
> all ones and all zeros, could not be used. I told them they were
> smoking
> something because I use them in our production network. They showed me
> in their study material that they state that they cannot be used. This
> floored me that they should be preaching this non-sense to network
> people. I never ran into a problem on our network and have never ran
> into this problem so I just wanted to know if it was legacy crap that
> is
> still being spread around or if there was an actual problem that could
> arise from using this subnets.
>
> IMHO this bad information should not be taught anymore. It's like
> saying
> that programs, on an Intel platform, have to reside in 64K of memory.
> This may have been true at one time but it is not the case anymore and
> they have stopped saying this a while ago. So we should say, it used to
> be considered bad practice but this is no longer the case.
>
> L8r
>
> -----Original Message-----
> From: Dan.Thorson@seagate.com [mailto:Dan.Thorson@seagate.com]
> Sent: Friday, August 09, 2002 1:41 PM
> To: Wade Edwards
> Cc: ccielab@groupstudy.com; Wright, Jeremy
> Subject: RE: ip addr 10.0.0.254/24
>
>
> Well, in your example of net-10 with a /24, the IP address
> 10.255.255.255
> could be either of the following
> 1) the all-subnets-directed broadcast (all subnets in the classfull
> network
> 10/8)
> or
> 2) the subnet-directed broadcast for network 10.255.255.0/24
>
> So the reason a subnet portion of all 1's is bad is pretty easy to see
> (which is it, all subnets, or just one?).
>
> On the other hand
> 10.0.0.0
> in one case is the network number for the class-full network 10/8, but
> is
> also the network number for the 10.0.0/24 subnet. This could lead to
> some
> routing confusion, I suppose, if you don't keep track of subnet
> masks....
>
> ... and as has been said previously, modern [Cisco] network equipment
> can
> keep track of the difference between these two routes (with ip
> subnet-zero
> enabled).
>
> danT
>
> ========================================
> Dan Thorson - Seagate Technology, LLC
> desk +1 (952) 402-8293 fax +1 (952) 402-1007
> SeaTel 8-402-8293
> ========================================
This archive was generated by hypermail 2.1.4 : Sat Sep 07 2002 - 19:48:23 GMT-3