DHCP, SuperScopes and Migrating Address Classes

A customer have a class 192.168.1.0/24 network with DHCP and wann "expand"
because the number os machnes in the environemtn is on the limit.
How can ths be done with minimal impact?

Creating another scope and using SuperScopes to distribute another class C
is an option to "expand"?

I´ll have to use a "router" (A linux or Windows doing routing) between the
networks?

"MCT Since 1999" <(E-Mail Removed)> wrote in message
news:#(E-Mail Removed)...
>
> A customer have a class 192.168.1.0/24 network with DHCP and wann "expand"
> because the number os machnes in the environemtn is on the limit.
> How can ths be done with minimal impact?
>
> Creating another scope and using SuperScopes to distribute another class C
> is an option to "expand"?
>
> I´ll have to use a "router" (A linux or Windows doing routing) between the
> networks?
>
You don't need any superscopes, just two 24-bit scopes (say
192.168.1.0/24 and 192.168.2.0/24).

If both scopes are on one DHCP server, you need a router between them
which supports DHCP relay, so that machines in the second subnet can receive
network config. A router will normally block DHCP broadcasts, but DHCP
relay allows the router to forward the request to the DHCP server in the
other subnet.

In this environment i suggested simply changing the subnet address of the
static IP Servers (30 of then), destroy DHCP scope and create another one
with Class "B" mask and make changes to the Router and Firewall/proxy
structure (Checkpoint and ISA)

Additionally, contact the Telco to check if thre is RIP active or if thres
is a need for manual intervention on the border routers to "know" that a new
network exists.


Currently there is a 192.168.1.x/24 net and everything will change to
192.168.x.y/16 and other Sites (locations) uses a 10.x network connected by
a MPLS VPN connection


"Bill Grant" <not.available@online> escreveu na mensagem
news:(E-Mail Removed)...
>
>
> "MCT Since 1999" <(E-Mail Removed)> wrote in message
> news:#(E-Mail Removed)...
>>
>> A customer have a class 192.168.1.0/24 network with DHCP and wann
>> "expand" because the number os machnes in the environemtn is on the
>> limit.
>> How can ths be done with minimal impact?
>>
>> Creating another scope and using SuperScopes to distribute another class
>> C is an option to "expand"?
>>
>> I´ll have to use a "router" (A linux or Windows doing routing) between
>> the networks?
>>
> You don't need any superscopes, just two 24-bit scopes (say
> 192.168.1.0/24 and 192.168.2.0/24).
>
> If both scopes are on one DHCP server, you need a router between them
> which supports DHCP relay, so that machines in the second subnet can
> receive network config. A router will normally block DHCP broadcasts,
> but DHCP relay allows the router to forward the request to the DHCP server
> in the other subnet.

"KayzerSoze" <(E-Mail Removed)> wrote in message
news:%(E-Mail Removed)...

> Currently there is a 192.168.1.x/24 net and everything will change to
> 192.168.x.y/16 and other Sites (locations) uses a 10.x network connected
> by a MPLS VPN connection

There are reasons why that is not done so already. Ethernet efficiency nose
dives after around 250-300 host. It may not be "humanly" noticable at
first,...but it exists. Obviously it is more noticable on a 10mbps LAN than
a 100mbps LAN,..and more noticable on a 100mbps LAN than a 1000mbps
LAN,...but if you don't want your 100mbps LAN to not run like a 10mbps
one,...or don't want your 1000mbps LAN to not run like a 100mbps one,...then
stick with the principle of 254 Hosts per segment.

A /16 bit mask creates 55,534 hosts is one segment. Who is going to have
that many hosts on a segment?? How many IP# are going to be wasted?...and
then possibly running into address conflicts when VPNing with some other LAN
in the future because you are eating up so many addresses needlessly.

If you use 10.0.0.0/16
Someone else uses 10.0.45.0/24
Attempt to VPN = Address conflict

The /24 bit mask is the perfect size with 254 hosts.

The lower bit masks are for supernetting over backbones where they are used
in router's routing tables to reduce the number of route table entries to
keep the tables efficient,...that is how the routers out in "Internet Land"
function. But the "large segments" are expected to be broken down into
smaller pieces (245 hosts or less) when they finally get down the "host
networks" where the individual PCs "live".

Follow the process that Bill explained.
Routers are not that expensive,...if they have the money to have enough
Hosts to crowd a "254 host" LAN then they have enough money to buy a LAN
Router or a Layer3 Switch (LAN Router and Switch built into same hardware
device).

--
Phillip Windell
www.wandtv.com

The views expressed, are my own and not those of my employer, or Microsoft,
or anyone else associated with me, including my cats.
-----------------------------------------------------