Hi. Hejc,
(2011/08/15 22:43), Nejc Škoberne wrote:
SA46T does not have address sharing mechanism. However, SA46T-AS have
address sharing mechanism. That is A+P like mechanism.
SA46T-AS is documented in
http://tools.ietf.org/html/draft-matsuhira-sa46t-as-01
In this version, two usages is discussed, one is clients environments,
and other is server environments.
As far as I can see from the SA46T-AS draft, you just embed 16-bit port
number in the SA46T-AS address, is this correct? This means, you cannot
have a port range, but only 1 port per CPE allocated. I guess I am
missing something here?
That's depend on prefix length. If prefix length is 128bits, only one
port per SA46T-AS tunnel endpoint is allocated. However, If prefix
length is 128 - 8 = 120bit, 2^8 = 256 ports per SA46T-AS tunnel endpoint
are allocated. And also, if prefix length is 128 - 16 =112bit, 2~16 =
65536 ports per SA46T-AS tunnel endpoint are allocated, and this case
dose not share one IP address.
Number of ports can be selected by prefix length.
Also, you say in the 5th section, that ICMPv4 doesn't work. But for
for this, you can make an ALG, like with other address sharing mechanisms?
I think having ICMPv4 work is quite a crucial thing in such a mechanism.
Yes, your comment is correct. I merely describe a general rule. And I
agree your point.
---
Question to others: what do you think about this? Should I also include
this one into the table of IPv4 address sharing mechanisms?
Thanks,
Nejc
_______________________________________________
Softwires mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/softwires
