Contradictions (Was: [Rfid] Re: XML vs. Text vs. Binary)

The other inherent contradiction that I see in these discussions is:

(1) We need to use a binary encoding because the latency of decoding 
XML would interfere with SLRRP's real-time requirements,

AND

(2) Our secure transport will the TLS over TCP/IP.

What is wrong with this picture?

This is why I keep getting back to the question of what the real-time 
requirements actually are.  It seems very unlikely to me that TLS 
over TCP/IP would meet the goals of a "real-time control protocol", 
but that parsing XML on top of that would push us over the edge...

I can't make a coherent argument on that front, though, because I see 
absolutely no real-time requirements.  What are the real-time 
requirements for a controller that says, effectively "Here are your 
inventory settings, run this until I say stop.  I'll be back later to 
pick up the data."?

Margaret

At 1:42 PM +0200 7/22/05, Stephane Bortzmeyer wrote:
>On Fri, Jul 22, 2005 at 06:56:14AM -0400,
>  Scott Barvick <sbarvick@revasystems.com> wrote
>  a message of 121 lines which said:
>
>>  the protocol processing is layered on top of standard security
>>  mechanisms as discussed in Section 4.2 of the draft.
>
>Section 4.2 is nice but it just says that TLS MUST be present, not
>that it MUST be used, no?
>
>>  Therefore, the implementation can safely access fields in payload as
>>  efficiently as possible.
>
>It seems to me quite contradictory to say (David Husak's message) "We
>will use fixed-length encoding because it is faster" and "security is
>not an issue, since we use TLS". Surely, TLS processing is much more
>costly than XML processing and therefore the cost of XML is
>negligible?
>
>
>
>_______________________________________________
>Rfid mailing list
>Rfid@lists.ietf.org
>https://www1.ietf.org/mailman/listinfo/rfid

_______________________________________________
Rfid mailing list
Rfid@lists.ietf.org
https://www1.ietf.org/mailman/listinfo/rfid