Le 3 août 2011 à 15:46, Simon Perreault a écrit : > On 2011-08-03 09:32, Rémi Després wrote: >>> I think there is an important point missing from this discussion. It is >>> tricky but it has important practical consequences. >>> >>> As I said, "The 900G figure is valid, *as long as internal hosts reuse >>> the same source address+port for different destinations*." >>> >>> The "as long as ..." part is important. >> >> Agreed. >> >>> I don't know of any operating >>> system that behaves like that. >> >> The point is that this behavior concerns the NAT44 of a CE that supports >> address sharing (it doesn't concern a PC OS). >> By permitting _this_ NAT to do endpoint-dependent mapping for TCP >> connections, the number of supported connections can be largely increased >> (if found useful). > Agreed.
> IMHO, a NAT should be allowed to do endpoint-dependent mapping > for protocols that are known not to cause issues (e.g. HTTP and DNS). Agreed. > I > think this would address most scaling concerns people have with port > ranges. Note that, even without that, expressed concerns on this tread are in general excessive because this thread only concerns port sets used for residual IPv4 traffic where IPv6 has been deployed. Numbers of needed ports will quickly decrease. Also, one should not forget that assigning full IPv4 addresses to DSL customers who need it remains possible with stateless solutions (presumably at a price, but we know there is no free lunch). Cheers, RD > This was already discussed in the behave WG by the way... > > Simon > -- > DTN made easy, lean, and smart --> http://postellation.viagenie.ca > NAT64/DNS64 open-source --> http://ecdysis.viagenie.ca > STUN/TURN server --> http://numb.viagenie.ca _______________________________________________ Softwires mailing list [email protected] https://www.ietf.org/mailman/listinfo/softwires
