Dean Anderson wrote: > On Thu, 22 Apr 2004, Tony Hain wrote: > > > You appear to overlook the case that H-D ratios apply to large complex > > enterprise networks just as they do to ISPs. Also, it is not necessary > for > > all nodes to need public access. As soon as any do there is a need to > avoid > > using any public prefixes on the internal network. > > I agree completely. But even large complex, enterprise networks don't need > complete, fully routed connectivity. As I said, I used to work for > Hitachi, which is a very large enterprise. This is why I doubt that > anyone really needs more than the RFC1918 space. > > > I understand what you wrote, but I think you are being overly > simplistic. In > > some scenarios it is very likely that only 5% of the nodes need public > > access. This creates a situation where acquiring more public allocation > is > > impossible due to current policy. At the same time there is no room to > grow > > without guessing which /8's are going to be allocated last. > > I suppose I am oversimplying. But I'm not sure I follow you. > > A large enterprise network has somewhat similar characteristics to a group > of companies operating over the public network. Only a small fraction of > the enterprise needs to get to every other part of the enterprise. > Enterprises can use NATs internally the same way other companies use NATs > with the public network. You assume a specific application and traffic model when you claim that enterprises can use NAT internally. This is not universally true. > > My point is that it takes a rather unusually extreme case to have an > enterprise that needs full connectivity internally, yet is so large that > it uses all of the available RFC1918 address space. There are only a few > possible organizations worldwide that could be in this group. Yes there are a small number, which is part of the reason they don't want to stand up and be bashed by the IETF. They just want to use the technology as it was intended to be used, and either need public space they will use for non-public nodes, or additional private space. > > There are some extreme cases of disconnection. For example, the internal > NSA networks are never going to be interconnected with the public network. > It would certainly be OK for the NSA to use address space other than > RFC1918 space. However, I don't think the NSA is large enough to exhaust > the RFC1918 space. Indeed, I doubt that Hitachi is large enough to exhaust > the RFC1918 space. I don't know about either of your examples, but basing need on historical use by traditional 'hosts' is ignoring the need to support growth and new applications & appliances. > > The main problem with RFC1918 space is inconvenient collisions. The same > space is used by 2 different divisions, that later need to interconnect. > It is not a lack of space--it is a lack of coordination. For companies > like Hitachi (not to pick on Hitachi), this would be just bad management. > They could have coordinated this. But in other cases, such as when > companies merge and find that they picked the same blocks, there is no way > they could have anticpated or coordinated the parts of RFC1918 space to > use. More space doesn't solve this problem either (which I suspect is the > real motivator, though it was not stated as such) Some organizations would no doubt be happy to have additional space to reduce some of their pain caused by churn coupled with the scarcity of IPv4 addresses. That is not the motivator for the draft I submitted. There are organizations that need more space now, and while they will be moving to IPv6, they can't deploy all the pieces fast enough. Tony _______________________________________________ Ietf@xxxxxxxx https://www1.ietf.org/mailman/listinfo/ietf