Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

My 10000+ users network, which spans the whole State and is very complex, has a "strange" addressing scheme.

Though our PCs are not directly connected/exposed to the Internet, our network designers assigned IP addresses taking them from a range different the "ordinary" IANA-reserved private IPv4 network ranges (10.0.0.0-10.255.255.255, 172.16.0.0-172.31.255.255, 192.168.0.0-192.168.255.255).

Assume that the IP addresses used in our intranet are in the range 20.*.*.* , i.e. addresses that are officially assigned in Internet (and don't belong to us).

Can anyone explain the advantages (if any) of this strange choice?

share|improve this question
add comment

2 Answers

up vote 25 down vote accepted

Don't do this if you ever intend to connect the network to the Internet. It's just far too risky.

First, you're using blocks of IP address space which belong to someone else. Because of this, you will have difficulty communicating with that other party as your routers may get confused as to whether the traffic should be sent to the other party or your internal network.

Along the router confusion line, this is a seriously non-default configuration, and the slightest mistake can result in live traffic with those IP addresses going over the public Internet, or worse, routes being announced to the Internet's default-free zone. Just like when somebody in Pakistan screwed up a router config and caused all of YouTube's traffic to be routed to that country, you could find yourself swamped with the other party's traffic.

And many ISPs and peering/transit providers have terms of service which prohibit using others' IP address blocks. If you use other people's IP address blocks, and they leak onto the Internet, you could be nullrouted or depeered or worse.

(Interestingly, Apple was one of the first companies to make this mistake; they had to renumber 5000 machines to recover. Their story is mentioned in RFC 1627.)

Since you or your predecessors already did it, your only way forward is to fix the numbering scheme. This is not particularly challenging technically, but it will be very time consuming and require some maintenance windows as well as coordination between the system and network administrators. Hopefully you can finish before something really bad happens.

share|improve this answer
1  
+1. don't do this, indeed. It means at least trouble to discuss with the whole Internet 20.X servers (which may not mean many "interresting" servers, but you can never be sure. For example, what if you need to access their website? or send them an email via your internal mailer?). There would be ways to circumvent those (ex: add specific 20.a.b.c/32 routes toward the Internet side, but would probably be set on every host and servers that needs to send packets to that adress). Iow, use the 10.x reserved ranges, and be happier (but it will take some work) –  Olivier Dulac Jul 5 '13 at 9:48
add comment

There are no advantages to this choice. The 20.0.0.0/8 network is not RFC 1918. Your network designers made a mistake and you are using someone else's public IPv4 addresses.

But it's surprisingly common among companies. Especially the 1.0.0.0/8 network. That one is abused a lot.

I foresee a big re-IP project in your company's future.

Edit: Covering the off-chance that you are CSC Corporation and you actually own 20.0.0.0/8, or were leased a block of that address space from them, then... yeah you're totally cool to use 20. IP addresses. :)

share|improve this answer
    
If the company owns that range, this is just wasteful. Upside of this though is that no NATing would be required, and in case of any merges/etc there won't be a problem with clashing addresses. –  Teftin Jul 4 '13 at 20:56
    
According to my quick Wikipedia search, unless he is CSC Corportation, he doesn't own 20.0.0.0/8. –  Ryan Ries Jul 4 '13 at 21:00
4  
I disagree. Prior to RFC 1918 everyone used publicly routable ip addresses internally. Not using RFC 1918 addresses isn't wrong or wasteful if the ip address space was designed and is managed properly. I work with a client that uses a publicly routable CIDR block internally and has 90% of those addresses in use. I see nothing wrong or wasteful about that. –  joeqwerty Jul 4 '13 at 21:05
2  
@joeqwerty in the times of IPv4 address exhaustion, keeping ip perfectly good public addresses for something which does not require them might be considered waste. –  Teftin Jul 4 '13 at 21:10
2  
Well, IPv4 address space was handed out wastefully in the beginning, but the internet was young then, and we were naïve. Regardless, I think we can agree that 4.2 billion IP addresses is not enough, no matter how they are distributed. Viva la IPv6! –  Ryan Ries Jul 4 '13 at 21:15
show 3 more comments

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.