Gordon Messmer <[EMAIL PROTECTED]> wrote:
>> Maybe they can run DHCP for
>> their entire network on 1 server with 1 IP address since there's no
>> physical separation anyhow.
>They could run DHCP like that, but the Linux box wouldn't be able to
>speak one of those subnets without a route to them, which there normally
>won't be without the aliased IP. That subnet would have trouble talking
>back, and the subnets can't talk to each other (even on the same
>physical segment) over TCP/IP.
I'm wondering if they have configured their router to forward broadcast
traffic across the subnets somehow. Again, I don't know much about their
router configuration.
>> For those PC's that should always be on a
>> specific subnet, they could use a host/MAC mapping which is what I do on my
>> network. Takes a little work to collect the MAC addresses from arp, but
>> once it's working it doesn't take much upkeep.
>Correct. Why are they subnetted anyway?
Beats the h*ll out of me. Partly to provide "security" of some kind
(although I don't know what), or so I'm told. Partly because they needed
to divide up their class B across several WAN links and partly because I
think whoever setup their network originally didn't *exactly* know what
they were doing. I think I'd have divided up the class B a little
differently, but hey, it's not my network...
>> That's kind of what I was wondering about. Fortunately, the aliased IP
>> address serves no other purpose than DHCP - and since DHCP is a broadcast
>> type service instead of source/destination (as you mentioned), I'm
>> wondering if the alias is needed at all.
>That depends. I pointed out why above. Exactly what functions does the
>Linux box serve?
The aliased IP provides only DHCP - nothing else. I *think* when all is
said and done the box will provide DHCP and DNS only.
>> The only thing I do know about their router is the routers for the virtual
>> subnets are also "virtual" since there's no physical separation. (Or as my
>> friend likes to call them "a router on a stick".)
>Yes, and if the networks aren't separated, that means that any
>communications between those subnets is transmitted on the same physical
>segment twice...Yuck!!
Yep. Again - not the way I'd have set things up.
>> That's *exactly* what I was wondering about yesterday. Since the network
>> segments are not physically separated how would DHCP know which subnet a
>> computer is supposed to be on - it wouldn't. What I'm thinking is DHCP
>> will assign addresses out of pool A until there's no more left and then
>> assign from pool B. Is this correct thinking?
>I don't know. I think that DHCP will simply server up all of the pool
>A, and then stop giving out new ones.
That makes two of us. I may have to experiment with that if I can convince
them it's in their best interest to find out.
>> Right. That's how I would have done it were it my network, but it isn't.
>> To be honest, I'm not really sure why their network is setup the way it is
>> - it seems a little goofy to me but apparently it works for them. To each
>> his own.
>It's a lot goofy. And it's inefficient. If they're using 10Mbit,
>they're going to run out of bandwidth fast.
Ok, I'll go with a lot goofy. ;-) Fortunately for them, they have an ATM
backbone.
>> No offense taken. You've already helped me to clarify a few things I was
>> wondering about - some of them before I had even asked the question.
<grin>
>Cool.
Yep!
-Eric
Eric Sisler
Library Computer Technician
Westminster Public Library
Westminster, CO, USA
[EMAIL PROTECTED]
Linux - don't fear the Penguin.
Want to know what we use Linux for?
Visit http://gromit.westminster.lib.co.us/linux
--
To unsubscribe: mail [EMAIL PROTECTED] with "unsubscribe"
as the Subject.
