Re: [ipfix] Slides from IETF presentation.

[Stewart Bryant <[e-mail]>]

OK, I am happy to change to this encoding.

I would imagine that when the IETF runs out of 2**15 IEs it
will register iteslf as a vendor giving it another 2**15 IEs :)

Stewart

MEYER,JEFFREY D (HP-Cupertino,ex1) wrote:

> Hi,
> 
>   In the interest of trying to borrow from existing IETF work.  My
> recommendation would be to go with Stewart and Ganesh's proposal in 4.2
> Compressed VI Qualified Template Flowset.
> 
>   http://www.ietf.org/internet-drafts/draft-bryant-ipfix-vendor-ie-00.txt
> 
>   This most closely resembles the way in which Diameter deals with vendor
> proprietary and IETF defined field (attribute) id's.
> 
>   See section 4.1 AVP Header in http://www.faqs.org/rfcs/rfc3588.html
> 
>   I would recommend that instead of having "reserved ranges" of id's,
> however that we reduce the size of field id's from 2**16 to 2**15 and
> specify the high order bit as a "Vendor" flag.  (Similar to the Diameter
> model)
> 
>   This gives us about 32,000 IETF defined fields which can be transmitted.
> Hopefully this will be enough, along with 2**32 * 2**15 additional namespace
> for proprietary fields.
> 
>   The diagram in 4.2 would be modified as:
> 
>     0                   1                   2                   3
>     0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
>    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
>    |V|       Field Type            |         Field Length          |
>    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
>    |                        Vendor-ID (opt)                        |
>    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
> 
>   Instead of talking about codepoint ranges for field types, you would just
> discuss whether the "V" bit was set.  If the V bit is set, then the
> Vendor-Id is present, otherwise it is not.
> 
>  
> Regards,
> 
>   Jeff Meyer
> 
> 
>>-----Original Message-----
>>From: majordomo listserver 
>>[mailto:[e-mail]]On Behalf
>>Of Ganesh Sadasivan
>>Sent: Tuesday, November 18, 2003 7:26 PM
>>To: Mark Fullmer
>>Cc: [e-mail]; [e-mail]
>>Subject: Re: [ipfix] Slides from IETF presentation.
>>
>>
>>
>>
>>Mark Fullmer wrote:
>>
>>
>>>Stuart, could you please post your slides detailing the options for
>>>extending the field ID space to include vendor extensions.
>>>
>>>If my memory is correct you proposed an optional 32 bit vendor
>>>ID which would either always be in the template or optionally
>>>in the template based on a "magic" field type which indicates
>>>a vendor extension. 
>>
>>Yes based on fixed ranges assigned to IETF defined and Vendor
>>defined  field types (some thing like {0-32767} for IETF defined
>>and {32768-65535} for vendor specified field types).
>>
>>
>>>
>>>After thinking about this for a few days I'd prefer the always
>>>there version of your proposal.  I don't see any reason to save
>>>a few bytes in the template at the expense of extra work in the
>>>decode/encode process. 
>>
>>When there is a mixture of IETF defined and  Vendor specified fields
>>and we are using large number of fields in the templates, 
>>this would be
>>useful. But otherwise there are'nt any major advantages.
>>I don't know how much extra work is involved in encoding/decoding.
>>Could be as simple as :
>>if (field_type < IETF_DEFINED_MAX) { ..} else {..}
>>
>>Or do you see something more worse?
>>
>>I am ok either way.
>>
>>
>>
>>>Also have you considered just using
>>>a 32 bit vs. 16 bit field type with say the first 16 bits reserved
>>>as a vendor ID?
>>
>>We just wanted to give a 48 bit id for the field type.
>>
>>Thanks
>>Ganesh
>>
>>
>>>mark
>>>
>>>
>>>-- 
>>>Help        mailto:[e-mail] and say "help" in 
>>>message body
>>>Unsubscribe mailto:[e-mail] and say
>>>"unsubscribe ipfix" in message body
>>>Archive     http://ipfix.doit.wisc.edu/archive/
>>>
>>
>>
>>
>>--
>>Help        mailto:[e-mail] and say "help" 
>>in message body
>>Unsubscribe mailto:[e-mail] and say
>>"unsubscribe ipfix" in message body
>>Archive     http://ipfix.doit.wisc.edu/archive/
>>
> 
> 
> 


--
Help        mailto:[e-mail] and say "help" in message body
Unsubscribe mailto:[e-mail] and say
"unsubscribe ipfix" in message body
Archive     http://ipfix.doit.wisc.edu/archive/