Re: [Softwires] PSID format as described in rfc 7597 vs. rfc 7598

[Xing Li <[e-mail]>]

[e-mail] 写道:
> Normen,
>
>   
>> I just want to make sure that I'm not misunderstanding this:
>>
>> RFC 7597, section 6 describes the last 16 Bits of the construction of the Ipv6 interface identifier as follows:
>>
>>    The PSID field is left-padded with zeros to create a 16-bit field.
>>
>> RFC7598 Section 5.1. describes the format for provision the PSID to clients with the following text:
>>
>>    o  PSID: 16 bits long.  The PSID value algorithmically identifies a set of ports assigned to a CE. The first k bits on the left of this field contain the PSID binary value.  The remaining (16 - k) bits on the right are padding zeros.
>>
>> To me this seems that the two RFCs use two different formats to express the same information in a field with the same name.
>>
>> Lets assume an example where the port split ratio is 6 (=k), slicing the IPv4 address up into up 2**6 = 64 slices, each segment having 1024 ports.
>>
>> In RFC 7597, to select the third port-range, the parameters become OFFSET 0, PSID-LEN 6, PSID 0x2 (left padded with zeros to 16 bits)
>> In RFC 7598, to select the third port-range, the parameters become OFFSET 0, PSID-LEN 6, PSID 0x800 (leftmost k-bits on this field contain the PSID binary value, which is right padded with zeros to fit 16 bits)
>>
>> So,
>> 1, Is my understanding of the two RFC's PSID formats correct?
>>     
>
> I believe so.
>
>   
>> 2, What's the reason for the difference in the formats?
>>     
>
> Good question. The 7597 PSID in the IID is mainly there for pretty printing / troubleshooting, and it makes sense to left pad it.
> I can only guess about the 7598 format, possibly to keep the PSID field consistent with the other fields (prefix) which are all right padded.
>   

I agree with Ole. For your information, our implementation has a 
configuration function to select the format of PSID. Regards, xing

> Best regards,
> Ole
>