I also agree this is a very nice piece of work. The current title is "Implementing AplusP in the provider's IPv6-only network". I think this not only covers "Implementing AplusP in an IPv6-only network", but also captures very important information about the port usage in various applications. I would like to see the draft to expand to include more popular applications and OS, and the radio between port usage and number of sessions. By reading Fig. 3 and Fig. 4 in http://opensourcev6transtechnologies.weebly.com/experiments-results.html, the radio is roughly 1 to 4. These experiments give concrete results to operators to plan for any port-sharing technology.
Cheers, Yiu On 8/10/11 8:38 AM, "Rajiv Asati (rajiva)" <[email protected]> wrote: >Xiaohong, > >Your work is very insightful. Thanks for pointing it out. > >In your experimentation, did you use 'consecutive' port-set or >'scattered' port-set or both? If just the former, then is it possible to >experiment using 'scattered' port-set? > >Cheers, >Rajiv > > >> -----Original Message----- >> From: [email protected] [mailto:xiaohong.deng@orange- >> ftgroup.com] >> Sent: Wednesday, August 10, 2011 5:35 AM >> To: [email protected]; Rajiv Asati (rajiva) >> Cc: [email protected] >> Subject: RE: [Softwires] Clarification of the stateles/stateful >>discussion >> >> >> >> |-----Original Message----- >> |From: Rémi Després [mailto:[email protected]] >> |Sent: Wednesday, August 03, 2011 4:02 PM >> |To: Rajiv Asati (rajiva) >> |Cc: [email protected] >> |Subject: Re: [Softwires] Clarification of the >> |stateles/stateful discussion >> | >> | >> |Le 3 août 2011 à 00:39, Rajiv Asati (rajiva) a écrit : >> | >> |> Satoru-san, >> |> >> |> This is an important point that most of us forget that >> |restricting to >> |> "n" ports doesn't equate to just "n" NAT sessions rather many more >> |> than n sessions. We must add that to the 4v6 motivation >> |draft as well >> |> as to the 4v6 comparison draft. >> | >> |+1 >> >> +1. Session number are usually many more than ports number. >> >> We tested port and sessions assumptions of applications on the A+P NAT, >> which has been documented in A+P experiments draft, to investigate how >> many ports and sessions they are costing on the NAT mapping table >>dynamically >> from the first NAT bindings being established to the last one being >> destroyed. >> >> Some results below: >> >> Usually, apps consume some more sessions than ports, and for the P2P >>apps, >> sessions >> number could even be as a couple of times as ports number. >> >> For example, BitTorrent established five hundreds of sessions while >>the port >> consumption >> was under a hundred in the first minute of the communication, because >>when >> BitTorrent >> initiates a downloading, it first uses the same source port to connect >>to the >> different destinations >> (destination IP and port) therefore one source port multiplexing >>different >> sessions. Skype is >> another example that uses one source port to multiplex different >>sessions >> thereby saving >> source port consumptions on NAT. >> >> For exact figures of ports/session number for apps, see my page: >> >>http://opensourcev6transtechnologies.weebly.com/experiments-results.html >> The session consumption comparison among the same set of applications is >> illustrated in Figure 4. >> >> Xiaohong >> >> >> | >> |RD >> | >> |> >> |>> port. There may be multiple session to different destinations using >> |> the same >> |>> external port. The 900G figure is valid, as long as internal hosts >> |> reuse the >> |>> same source address+port for different destinations. >> |> >> |> Cheers, >> |> Rajiv >> |> >> |> >> |>> -----Original Message----- >> |>> From: [email protected] [mailto:[email protected]] >> |> On Behalf >> |>> Of Simon Perreault >> |>> Sent: Tuesday, August 02, 2011 9:55 AM >> |>> To: [email protected] >> |>> Subject: Re: [Softwires] Clarification of the stateles/stateful >> |> discussion >> |>> >> |>> Simon Perreault wrote, on 08/02/2011 09:24 AM: >> |>>> Satoru Matsushima wrote, on 08/01/2011 10:41 PM: >> |>>>> Thanks, a clarification has made to clear a confusion of >> |restricted >> |> port >> |>>>> set/ranges and NAT session table limitation. Even if a CPE is >> |> allocated 256 >> |>>>> ports, NAT session can be made over 900G sessions in theory. >> |> ('2^32'<Full >> |>>>> 32bits v4 address> - '2^29'<class-D/E> - '2^7'<0/8,127/8>) * >> |> 2^8<256 >> |>> ports>. >> |>>> >> |>>> This is only true for endpoint-dependent NATs, which are forbidden >> |> by the >> |>> BEHAVE >> |>>> RFCs (4787 and 5382). According to these RFCs, NATs MUST have >> |>>> endpoint-independent mapping behaviour. This means that each NAT >> |> session >> |>> will >> |>>> consume one external port. >> |>> >> |>> Sorry, I confused the terminology. Each NAT *binding* will consume >> |>> one external port. There may be multiple session to different >> |>> destinations using >> |> the same >> |>> external port. The 900G figure is valid, as long as internal hosts >> |> reuse the >> |>> same source address+port for different destinations. >> |>> >> |>> 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 >> |> _______________________________________________ >> |> Softwires mailing list >> |> [email protected] >> |> https://www.ietf.org/mailman/listinfo/softwires >> | >> | >> | >_______________________________________________ >Softwires mailing list >[email protected] >https://www.ietf.org/mailman/listinfo/softwires _______________________________________________ Softwires mailing list [email protected] https://www.ietf.org/mailman/listinfo/softwires
