I don't understand why we care.  There are hundreds if not thousands of 
companies developing proprietary solutions with 15.4 radios that will not 
conform with 6lowpan, zigbee, HART, or any other 15.4-based wireless standard.  
Perhaps more importantly, the crc16 fcs is not sufficient to guarantee the 
integrity of received frames - roughly one out of a million will be corrupted 
in major ways but still bear a valid fcs.
 
So what are we hoping that these bits in the dispatch will do for us?  No 
matter how well we coordinate between standards, I guarantee that our radios 
will receive the occasional bogus packet with a "correct" dispatch byte and 
length.  You can try to type-check the fields all you want, but occasionally 
you will execute on a payload that is at least partly pure noise.  One 
alternative is to build extra redundancy into the headers, but header 
compression moves us in the other direction: if we do our job right we squeeze 
all of that redundancy out by design.
 
The simplest way to deal with this problem is to use a MAC layer MIC-32 in 
conjunction with the fcs.  If we build secure networks, we need that anyway. If 
we want an "open" network with no security, then we can define a default 
(semi-random) MAC-MIC key that all 6lowpan devices will use.  For most 
applications this reduces the probability of undetected errors to an acceptably 
low level.
The alternative is to have our motes periodically do unpredictable things.  
This is fine in a one hour demo or a one month pilot, but is generally frowned 
on in production.
 
ksjp

________________________________

From: Anthony Schoofs [mailto:[EMAIL PROTECTED]
Sent: Wed 5/23/2007 7:31 AM
To: 6lowpan
Subject: [6lowpan] Dispatch value bit pattern for 6lowPAN



Dear all, 

In the "Transmission of IPv6 Packets over IEEE 802.15.4 Networks" Internet 
Draft, section Dispatch type and Header page 7, is mentionned that  ".. other 
non-LowPAN protocols that wish to coexist with LowPAN nodes should include a 
byte matching this pattern (00 xxxxxx) ..". 

I had a look at the ZigBee network header and especially at the Frame Type 
sub-field (the 3 first bits after the 802.15.4 MAC header). 
For acknowledgement and command frames, the Frame Type value is either 010 or 
011 (data frames start with 00). 

Consequently, it might happen that you get a match with a LoWPAN pattern, and 
then a problem of coexistence between LowPAN nodes and ZigBee nodes. 

I would like to know if this coexistence issue was raised before and if it is a 
matter for the 6LoWPAN WG. 

Best regards, 
Anthony 

--
Anthony Schoofs
Philips Research
High Tech Campus 37 , 5656 AE Eindhoven, The Netherlands
Email: [EMAIL PROTECTED] 



_______________________________________________
6lowpan mailing list
[email protected]
https://www1.ietf.org/mailman/listinfo/6lowpan

Reply via email to