From: Jim Graves (jtg@xxxxxxxxxx)
Date: Thu May 17 2001 - 12:49:56 GMT-3
This doesn't sound like a terribly efficient design.
In general, using other people's addresses is Bad. Sure, today your boss
says this network will never connect to anything in the public
Internet. But imagine your living hell in two months, when some dude in a
suit decides it would be really keen for all these devices to be able to
hit some web site that's suddenly become mission critical.
Here's a few possible ways you could do this without stepping on anyone
else's address space:
* Abandon the cookie-cutter approach and allocate network space from
10.0.0.0/8 as needed. Why should a customer with 50 networks get the same
space as one with 3000?
* Embrace NAT. Even if you end up with more than 16 million end devices,
they probably won't have to speak to each other at the same time. You
could give each customer 192.168.0.0/16, and use a combination of static
and dynamic NAT to give them whatever connectivity is needed to the rest of
the network. Certain applications may have problems with this approach,
though.
* Become an early adopter of IPv6.
But the short answer to your question (as you've no doubt gathered by the
flood of responses), is: no, using someone else's IP space isn't considered
good design, but you're probably only hurting yourself if you do.
Jim
At 10:31 AM 5/17/2001 -0400, Bruce Williams wrote:
>I need enough addresses to accommodate hundreds of customers, each will need
>hundred or so different networks and in these networks they will have over a
>hundred hosts. I might be able to accommodate that by subnetting 10.0.0.0
>and then if necessary subnetting 172.16.0.0 and then 192.168.0.0, but my
>boss believes it will be easier to use Class A space 1.0.0.0 to 126.0.0.0
>that way they can have a cookie cutter method of assigning IP addresses.
>
>----- Original Message -----
>From: <andrew.2.shore@bt.com>
>To: <bruce@williamsnetworking.com>; <David.Halaska@getronics.com>;
><ccielab@groupstudy.com>
>Sent: Thursday, May 17, 2001 10:20 AM
>Subject: RE: Using Public Addresses Internally
>
>
> > ?????? ever thought of subnetting ?
> >
> >
> > Andrew Shore
> > BTcd
> > IDC Technical Specialist
> > > ignite Content Hosting
> >
> >
> > -----Original Message-----
> > From: Bruce Williams [mailto:bruce@williamsnetworking.com]
> > Sent: 17 May 2001 15:12
> > To: Halaska, David; ccielab
> > Subject: Re: Using Public Addresses Internally
> >
> >
> > We need much more address space than that.
> >
> >
> > ----- Original Message -----
> > From: "Halaska, David" <David.Halaska@getronics.com>
> > To: "'Bruce Williams'" <bruce@williamsnetworking.com>;
> > <ccielab@groupstudy.com>
> > Sent: Thursday, May 17, 2001 10:16 AM
> > Subject: RE: Using Public Addresses Internally
> >
> >
> > > Could you use the 10.x.x.x class A range that is reserved for private
>use?
> > >
> > > -----Original Message-----
> > > From: Bruce Williams [mailto:bruce@williamsnetworking.com]
> > > Sent: Thursday, May 17, 2001 8:56 AM
> > > To: ccielab@groupstudy.com
> > > Subject: Using Public Addresses Internally
> > >
> > >
> > > My company wants to use public addresses from the Class A range
> > internally.
> > > I
> > > realize the danger if these routes got advertised on the Internet, but
>is
> > > this
> > > something that is considered acceptable if it is carefully done to
>prevent
> > > the
> > > risk of these routes being propagated out on the Public Internet? These
> > > networks will be used to address equipment in a multitude of cellular
> > radio
> > > base stations around the country and they will only be connected to our
> > > network. There will central locations where users from the internet
>could
> > > access a database which will query these systems, but there will not be
>a
> > > direct internet connection. I would appreciate any advice on this.
> > >
> > > Thanks,
> > >
> > > Bruce Williams
> > > 215-275-2723
> > > bruce@williamsnetworking.com
> > > **Please read:http://www.groupstudy.com/list/posting.html
> > **Please read:http://www.groupstudy.com/list/posting.html
>**Please read:http://www.groupstudy.com/list/posting.html
**Please read:http://www.groupstudy.com/list/posting.html
This archive was generated by hypermail 2.1.4 : Thu Jun 13 2002 - 10:30:43 GMT-3