RE: Question on extended RADIUS attribute format in draft-congdon-radext-ieee802-02.txt

["Glen Zorn \(gwz\)" <gwz@cisco.com>]

Bernard Aboba <mailto:aboba@internaut.com> supposedly scribbled:

> It would be nice to allow larger attributes if there was some way
to
> accomplish this while maintaining backward compatibility.  Do we
have
> a proposal for how to do this?  

Beats me -- thought that this was one?

> 
> For example, if an attribute is defined with a length > 253, won't
> existing proxies have trouble figuring this out?  Or is the idea
to
> have the *inner* length different from the *outer* length?  That
way
> the outer length <= 253, but the inner length can be larger? 

That was my assumption.  Otherwise, why allocate 16 bits for
Length2?  The draft says (in Appendix a.2) 

In order to enable Extended attributes to be encoded more
economically, a "Short Form" of the Extended attribute format is
proposed.  The Short Form can be used to encode any Diameter AVP
that meets the following constraints: 
         [a] An IETF standard attribute (not Vendor-Specific) 
         [b] Diameter Code between 0 and 16384 (all existing
attributes) 
         [c] No flag bits other than the Mandatory bit.  
 
Aside from the fact that Length2 is only 2/3 the size of the
Diameter AVP Length field (thus making the assertion invalid
anyway), the intent seems to be to carry payloads larger than 253
octets.

> 
> Inquiring minds want to know :)
> 
> On Thu, 10 Feb 2005, Glen Zorn (gwz) wrote:
> 
>> Glen Zorn (gwz) <mailto:gwz@cisco.com> supposedly scribbled:
>> 
>>> It seems as if one of the purposes of defining the extended
format
>>> might be to allow larger attribute (i.e., > 253 bytes of data).
>>> However, the Length field of the extended format is defined to
>>> include the length of Type, Length, S, Code, Length2, Flags,
>>> Vendor-Id and Data fields.  If it includes the length of the
Data
>>> field then that would seem to limit the total length of the
>>> attribute to 253 octets anyway.  Am I just missing something, or
>>> should it say 
>>> 
>>> 	Length
>>> 
>>>       The length of the extended attribute header in octets,
>>>       including the Type, Length, S, Code, and Length2 fields.
>>> 
>>> With the definition of Length2 changed to
>>> 
>>>    Length2
>>> 
>>>       The length of the sub-attribute in octets, including the
>>> 	Length2, Flags, Vendor-Id and Data fields.
>>> 
>>> Or something similar?
>> 
>> BTW, if you do this, I believe that you can get rid of the 'S'
bit
>> (which seems a little weird to me) and keep the Flags field in
the
>> "Short form" at the cost of 6 bits.  The short and long forms
could
>> be distinguished by inspection of the Length field: 6 for short
>> form, 8 for long form.  In addition, you could carry 16 and 32
bits
>> of the Diameter type code in the short & long forms,
respectively.
>> 
>>> 
>>> ~gwz
>>> 
>>> Treat the Earth well.
>>> It was not given to you by your parents.
>>> It was loaned to you by your children.
>>>   -- Kenyan Proverb
>>> 
>>> Humankind has not woven the web of life.
>>> We are but one thread within it.
>>> Whatever we do to the web, we do to ourselves.
>>> All things are bound together.
>>> All things connect.
>>>   -- Chief Seattle
>> 
>> Hope this helps,
>> 
>> ~gwz
>> 
>> Why is it that most of the world's problems can't be solved by
simply
>>   listening to John Coltrane? -- Henry Gabriel

Hope this helps,

~gwz

Why is it that most of the world's problems can't be solved by
simply
  listening to John Coltrane? -- Henry Gabriel

--
to unsubscribe send a message to radiusext-request@ops.ietf.org with
the word 'unsubscribe' in a single line as the message text body.
archive: <http://psg.com/lists/radiusext/>