Yiu, Thank you for your good comments, even in the hurricane. I am proposing modify the texts in the document to be more accurate since PMTU is unable to avoid fragmentation here. Cache in mAFTR maybe one possible way, but I agree that it is more complicated. For the time being, there are two issues I can see. First, the path of multicast may be not consistent with unicast. Second, if m AFTR sends PMTU to every B4s and take the least common denominator, the efficiency will be reduced. There may be malicious reduction of PMTU value. But if we set the acceptable minimum value of PMTU at mAFTR, this problem may be solved.
Best Regards, Tina TSOU http://tinatsou.weebly.com/contact.html From: [email protected] [mailto:[email protected]] On Behalf Of Lee, Yiu Sent: Monday, August 29, 2011 12:16 AM To: [email protected] Subject: Re: [Softwires] Comments on section 6.3 of draft-qin-softwire-dslite-multicast-04 Hi Tina, To use PMTU in this scenario, this is more complicated than what you explain. First, this traffic flow is from mAFTR to the mB4. In between, there could be more than one IPv6 multicast routers. In our spec, mAFTR will replicate the packet once in the multicast I/F to the IPv6 MDT, can you explain to me your poropose how PMTU can work here? Does it mean mAFTR will send PMTU to every B4s and take the least common denominator? Regards, Yiu From: Jacni Qin <[email protected]<mailto:[email protected]>> Date: Sat, 27 Aug 2011 09:48:12 +0800 To: Tina TSOU <[email protected]<mailto:[email protected]>> Cc: "[email protected]<mailto:[email protected]>" <[email protected]<mailto:[email protected]>> Subject: Re: [Softwires] Comments on section 6.3 of draft-qin-softwire-dslite-multicast-04 hi, On Sat, Aug 27, 2011 at 2:32 AM, Tina TSOU <[email protected]<mailto:[email protected]>> wrote: Dear Jacni, Just after reading RFC 1981, I think fragmentation of IPv6 is needed. In section 5.1, it says, "It is possible that a packetization layer, perhaps a UDP application outside the kernel, is unable to change the size of messages it sends. This may result in a packet size that exceeds the Path MTU. To accommodate such situations, IPv6 defines a mechanism that allows large payloads to be divided into fragments, with each fragment sent in a separate packet (see [IPv6-SPEC] section "Fragment Header")." If the node which makes PMTU is the multicast source, it can change the size of message when the size exceeds the PMTU value. However, mAFTR, as a router, is unable to change the size of the message. Here is an example: An IPv4 packet is sent from the multicast source to mAFTR with size of 1000, while the IPv6 PMTU for mAFTR and mB4 is 960, how does mAFTR forward the packet? One possible way is to make IPv6 fragmentation, with two fragments: one is 960 and the other is 40 with "IPv6 Fragment Header". In addition, there may be also other ways to avoid fragmentation, e.g., cache in mAFTR as defined in [draft-jiang-behave-v4v6mc-proxy]. Jacni>: Please see Yiu's comments, while for the discovery if you really want to do it, this way to help you to understand, it's the end point of tunnel which seems to "source" the v6 packets. Hope you can get it. Cheers, Jacni _______________________________________________ Softwires mailing list [email protected]<mailto:[email protected]> https://www.ietf.org/mailman/listinfo/softwires
_______________________________________________ Softwires mailing list [email protected] https://www.ietf.org/mailman/listinfo/softwires
