1

I have a Windows Azure host that has an internal IP address of 100.x.x.x.

I would assume that this is conflicting with external hosts with the same net mask.

Is there any assurance that I will not experience a production IP conflict when Azure uses IP addresses that aren't "private"?

2 Answers 2

4

You get the same assurance as everyone else who has an IP address. If Azure has spare IP addresses it can use them internally or externally as it pleases, there is no requirement to use RFC1918 addressing internally.

8
  • so that is under the assumption that 100.x.x.x is owned by Microsoft or their subsidiaries? Jul 30, 2015 at 7:20
  • It's kind of assuming that the Microsoft network admins aren't clueless idiots that would just use random IP addresses that they have no control over. What's the number after 100 btw ?
    – user9517
    Jul 30, 2015 at 7:24
  • 100.72.44.x is the class C subnet Jul 30, 2015 at 7:28
  • 1
    That's part of the 100.64/10 allocation block reserved for RFC6598 IANA-Reserved IPv4 Prefix for Shared Address Space AKA Carrier Grade NAT. That doesn't really change the answer though.
    – user9517
    Jul 30, 2015 at 7:36
  • 1
    I must have run out of toes ;)
    – user9517
    Jul 30, 2015 at 9:09
4

100.0.0.0/8 was a allocated to ARIN. ARIN has allocated one quarter of this to be used as shared address space. The other three quarters have been allocated to various providers including T-Mobile, Verizon, Sprint, Wayport.

100.64.0.0/10 is shared address space to be used similarly to RFC1918 address space. The most significant difference is that you are only allowed to use 100.64.0.0/10 addresses on your LAN, if your router can automatically resolve any addressing conflicts due to addresses being duplicated on the inside and outside of your router.

2
  • do routers actually do that kind of conflict resolution? What kinds? New to me... never heard of it, but very curious about it. Probably Cisco can... Jul 30, 2015 at 8:02
  • @LamonteCristo Most routers cannot. That is the reason 100.64.0.0/10 got allocated in the first place. If routers had always been able to automatically avoid or mitigate such conflicts, then all CGN deployments would have been using RFC1918 address space. But with the introduction of RFC6598 it is now permitted for a router to use 100.64.0.0/10 on the LAN side as long as it can deal with such conflicts. By using 100.64.0.0/10 on the LAN side you avoid conflicts with routers even closer to the user still using RFC1918 address space.
    – kasperd
    Jul 30, 2015 at 9:01

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .