Dear Med, dear all,

I carefully checked France Telecom's IPR statement we just received about 4rd.
For reasons detailed below, I found that none of the three listed patents 
applies to 4rd.

Could you please, Med, check the technical analysis and, if we reach common 
understanding, see to it that the IPR statement be either deleted or 
appropriately updated. 

Thanks,
RD


__________________
TECHNICAL ANALYSIS

1. 4rd isn't concerned with patent EP 2294798 because uses of for port numbers 
for address sharing are radically different.
  
 - In EP 2294798, deriving a destination "identifier" from an original address 
+ port must include these two steps: (1) find which PORT RANGE is matched by 
the destination port among those that are recorded as relevant to this 
destination address; (2) take as destination identifier that which is recorded 
as associated with this address/port-range couple.
  
 - In 4rd, deriving a destination IPv6 address from an IPv4 A+P includes these 
two steps: (1) find which IPV4 PREFIX is matched by the IPv4 address, among 
recorded Rule IPv4 prefixes; (2) take as prefix of the IPv6 destination the 
concatenation of three parts: the Rule IPv6 prefix of the mapping rule that has 
the found IPv4 prefix; all bits that, in the IPv4 address, follow this IPv4 
prefix; some bits of the destination port. 

Looking for a match between an IPv4 address and an IPv4 PREFIX is radically 
different from looking for a PORT RANGE that contains a given port.


(Incidentally, the same analysis applies AFAIK to MAP-E and MAP-T.) 



2. 4rd isn't concerned with patents EP 2297927 and/or EP 2297927 because 
address formats are beyond doubt incompatible. 

- In both EP 2297927 and EP 2297928, each IPv6 address constructed from an IPv4 
A+P MUST CONTAIN the FULL IPv4 address FOLLOWED BY the port number.
 
- In 4rd, an IPv6 address derived from an IPv4 A+P NEVER contains the FULL IPv4 
address FOLLOWED by a port number:

 . The full IPv4 address is always embedded in bits 80-111 of IPv6 addresses, 
but always followed by a CNP field which has no relation to the port number 
(Figure 5).

 . With a CE mapping rule, the first 64 bits of the IPv6 address may contain 
part of the IPv4 address but NEVER the full address. (A CE mapping rule has a 
Rule-IPv4-prefix length k > 0  because with k = 0 it couldn't be distinguished, 
by address longest match, from the BR mapping rule whose k must be 0. With k > 
0, at most 32 - k bits of IPv4 addresses can be embedded in 4rd IPv6 prefixes.) 
   
 . With the BR mapping rule, no IPv4-address bit is embedded in the first 80 
bits of IPv6 addresses (Figure 5).


(Incidentally, the same reasoning doesn't apply to MAP-E and MAP-T because 
their PSIDs are repeated after embedded IPv4 addresses. Where PSID lengths are 
16 bits, which is permitted, IPv6 addresses do contain IPv4 addresses followed 
by port numbers.)
__________________







2013-04-05 19:53, IETF Secretariat <[email protected]> :

> 
> Dear Remi Despres, Reinaldo Penno, Yiu Lee, Gang Chen, Sheng Jiang, Maoke 
> Chen:
> 
> An IPR disclosure that pertains to your Internet-Draft entitled "IPv4 Residual
> Deployment via IPv6 - a Stateless Solution (4rd)" (draft-ietf-softwire-4rd) 
> was
> submitted to the IETF Secretariat on 2013-03-20 and has been posted on the 
> "IETF
> Page of Intellectual Property Rights Disclosures"
> (https://datatracker.ietf.org/ipr/2050/). The title of the IPR disclosure is
> "France Telecom's Statement about IPR related to 
> draft-ietf-softwire-4rd-04."");
> 
> The IETF Secretariat
> 
_______________________________________________
Softwires mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/softwires

Reply via email to