Hi Ben

>Perhaps this is a "newbee" question, answered in the history of this
>discussion thread, but how does a lower layer "know" what kind of MIC
may be
>applied by a higher layer (i.e. in the payload) unless it is explicitly
told
>by the next higher layer?  The suggested wording requires such cross
layer
>knowledge.  I see that only the source of the packet can know this, but
>likewise in that source node, how would the LoWPAN layer make such a
>determination? The obvious way is when informed by the upper layer via
the
>appropriate SAP mechanisms. Would the wording not be more 'crisp' if it
the
>only case is when the suppression of the UDP CRC is indicated by the
upper
>layer?

[Pascal] That's right. But for tunnels which are similar yet a bit
different.

>Mention is made of this being a layer violation in the thread, but what
I
>don't understand is why this is necessary in the standard?
[Pascal] 
Saving bits is critical in this space, mostly sensors->backbone because
that's most of the traffic and backbone routers have a power source
anyway. Bits translate into days and weeks of battery operation. Bits
can make the difference between a packet that can be sent on scavenged
energy and a packet that can not be. 2 bytes are wealth in LoWPANs :)

Pascal

>----- Original Message -----
>From: "Pascal Thubert (pthubert)" <[EMAIL PROTECTED]>
>To: "6lowpan" <[email protected]>
>Sent: Thursday, November 20, 2008 2:46 AM
>Subject: [6lowpan] Proposed new text for the UDP checksum elide issue
>
>
>> Hi:
>>
>> Carsten asked for a crisper wording for which node can elide the UDP
>> checksum and when that happens. There seems to be a consensus that
>> we are not ready to open the door for heuristics based operations.
>>
>> So at this point, only the source of a packet MAY elide the checksum,
>> And the receiver of a packet MAY ignore the checksum if it is
present.
>> This in only allowed in the case of tunneling or stronger integrity
>> check above UDP.
>>
>> Here is a proposed text:
>>
>> "
>>   With this specification, the source of a packet MAY elide the UDP
>>   checksum in the following cases:
>>
>>   Tunneling:  The source of the packet is tunneling another packet
that
>>      possesses its own integrity mechanism.
>>   Upper Layer MIC:  The Upper Layer Protocol over UDP uses an end-to-
>>      end Message Integrity Check (MIC) that has stronger properties
>>      than what is provided by the UDP checksum.  Such an integrity
>>      check MUST be end-to-end and cover the IPv6 pseudo-header, UDP
>>      header, and UDP payload.
>>
>>   Only the source of a packet can know what Upper Layer operation
takes
>>   place.  A router on the way that is not aware of such operation
>>   SHOULD NOT elide the UDP checksum when performing 6LoWPAN
>>   compression.
>>
>>   The 6LoWPAN termination point has to recompute the elided checksum
>>   based on the received packet.  If that point is a router, then the
>>   router reforms a proper IPv6/UDP packet that can be forwarded on
any
>>   interface.  If the 6LoWPAN termination point is the destination of
>>   the packet and is aware of the presence of the Upper Layer MIC for
>>   the destination UDP port, it MAY omit the UDP checksum operation
>>   completely.
>> "
>>
>> What do you think?
>>
>> Pascal
>>
>>
>> _______________________________________________
>> 6lowpan mailing list
>> [email protected]
>> https://www.ietf.org/mailman/listinfo/6lowpan
>>
>
>_______________________________________________
>6lowpan mailing list
>[email protected]
>https://www.ietf.org/mailman/listinfo/6lowpan
_______________________________________________
6lowpan mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/6lowpan

Reply via email to