I see 2 problems off the top of my head with using public IP blocks for private networks. 1) You're not going to be able to reach servers/services/etc that actually have allocated those IP blocks. (May or may not affect you, but that's your issue to deal with in the future). 2) (and more important) It really makes it easy to 'accidentally' announce that public IP block out in the future, unless you have proper announce filters in place (And if something as basic as subnetting isn't done properly, I doubt route filtering is either). This one not only affects you, but affects the netblock that gets mistakenly announced out. RFC1918 space was designed to prevent these issues. Ken Matlock Network Analyst Exempla Healthcare (303) 467-4671 matlockk@exempla.org -----Original Message----- From: sthaug@nethelp.no [mailto:sthaug@nethelp.no] Sent: Monday, February 02, 2009 11:56 AM To: darcy@druid.net Cc: nanog@nanog.org Subject: Re: Private use of non-RFC1918 IP space
Company A uses public IP block A internally. Company B uses public IP
OK, so we start out with a bad network design then.
No. We start with blocks A and B which are both properly allocated by the relevant addressing authorities.
block B internally. Company A and B later merge, and connect their networks. No conflict, no renumbering needed (at least not right away).
Maybe. What if they both happened to choose 1.2.3.4/8? Is this just a matter of decreasing the odds of a conflict? It still seems like bad network management to me.
My assumption throughout this whole discussion, which clearly has not been understood, is that the public IP block used internally is a properly allocated by the relevant addressing authority. That is, for me, the whole point of using public addresses to guarantee uniqueness. Steinar Haug, Nethelp consulting, sthaug@nethelp.no