Sites and Services/Subnets question

  • Thread starter Thread starter Jim Geith
  • Start date Start date
J

Jim Geith

I had a discussion with someone over proper subnet entries in Sites and
Services. This came up because he is using a private IP scheme on his
clients and servers with 10.100.2.x/24 subnetting. On his sites and
services he entered a 10.100.0.0/16 subnet and said it would contain all of
the subnets on that site. I disagreed based on getting them to route that
way on his Cisco network, and keeping AD straight were different, and that
AD would think those subnets do not belong to that site. If he is right it
could simplify things. I haven't been able to find a good definition except
for the many examples which conflict with his theory. Ideas?
 
Jim,

Technically your friend is right, the /16 actually contain
any subnet in the /24 if you look at proper subnetting,
what your friend did actually called suppernetting where
you minimize your routing table by cutting the subnet, by
doing that the router or the server will need less memory
in use and less cpu.

What ever belong to 10.100.0.0 will be forward to a
certain location, instead of looking at the individual
subnets.

The problem will be if you use some thing like that with a
public ip address, when you are not the person that own
all the subnets, then you will create bad routing decision
 
clients and servers with 10.100.2.x/24 subnetting. On his sites and
services he entered a 10.100.0.0/16 subnet and said it would contain all of
the subnets on that site. I disagreed based on getting them to route that
way on his Cisco network, and keeping AD straight were different, and that
AD would think those subnets do not belong to that site. If he is right it
could simplify things. I haven't been able to find a good definition except
for the many examples which conflict with his theory. Ideas?

From what you say your friend is right -- it's difficult to even
understand what you wanted to do differently.

10.100.0.0/16 says that every address between 10.100.0.0 to
10.100.255.255 is in that site.

You can even (as in routing) add 10.100.0.0/24 for a different site
and then only this latter range will be in the second site while
everything else stays in the first site. (i.e., 10.100.1.0 through
10.100.255.255 are still in first site while 10.100.0.0 through
10.100.1.255 are in the second.)
 
I couldn't remember the supernetting term. At any rate, the bottom line
question is whether AD supports the supernetting configuration properly, and
both of you said yes. Any reference material form MS I have read said the
IP/mask assigned in Sites and services must match the clients and servers.
No mention of using supernetting when possible. It would definitely simply
things. Our IP network is mostly structured with set consecutive subnet
ranges assigned to given geographic areas. We still have some places going
through the conversion to private IP space.

Thanks.
 
Well, usually it's called "route summary" or "net summarization"
when you are not actually "joining nets" but if you say 'supernetting'
we'll know what you mean, usually that term is for when someone,
e.g., an ISP actually joins multiple networks together.

Term doesn't really matter as long as both parties understand.

It's just a DEFINITION of subnets, as long as you describe
which addresses are in each site you may do it as you wish
but the summarization technique is most efficient.
 
Hmm if you could point out this documentation, we should go look into getting it changed. As my fellow MVP Dave Shaw
pointed out, Routing and Subnetting in AD have no bearing on each other. You can supernet AD subnets for sites all day
and in fact for any large deployments it is generally a rather common scenario.
 
Back
Top