Hi Ralph,

Please ignore the previous message, prematurely sent with no substance in it 
(apologies).

Le 12 août 2010 à 00:00, Ralph Droms a écrit :

>> 1. The text should take into account that "over time" more and more servers 
>> will have IPv6, and that the full 64K ports are available in IPv6. (The 
>> mentioned AJAX/Web 2.0 sites would typically enable IPv6 asap, if not done 
>> yet. This will prevent problems over time.)
> 
> traffic to IPv6 servers just bypasses dual-stack lite altogether, right, so 
> the limitation doesn't apply to that traffic?

Right.
Using IPv6 where possible is therefore the best approach against port shortage 
in IPv4.

>> 2. If the number of assignable IPv4 addresses is for a start multiplied by 
>> 10, by statically sharing ports of each address among 10 customers, this 
>> still leaves several thousands of IPv4 ports per customer. (Exactly 6144 
>> ports per customer if, as appropriate, the first 4K ports, that include 
>> well-known ports and have special value are excluded). 
> 
> Agreed; one could argue that even sharing an IPv4 address among 5 customers 
> allows 5x as many customers in the existing IPv4 address assignment, which 
> should be more than enough to bridge the gap until IPv6 is available.
> 
>> 3. Where applicable static sharing is much simpler to operate.
> 
> Agreed.

Apparently, this point should be better documented in IETF.
Would draft-ietf-intarea-shared-addressing-issues-01 be a good place for that?

>> If the principle is agreed, I can propose detailed words myself to improve 
>> the draft.

Med has independently suggested to just delete the section on static vs dynamic 
port allocations from the DS-lite draft.
This would be sufficient to cover my point in this draft.

Regards,
RD


_______________________________________________
Softwires mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/softwires

Reply via email to