On Jan 14, 2009, at 9:04 AM, Razvan Musaloiu-E. wrote:

> Hi!
>
> On Wed, 14 Jan 2009, Miklos Maroti wrote:
>
>> On Wed, Jan 14, 2009 at 4:47 PM, Razvan Musaloiu-E. <[email protected] 
>> > wrote:
>>> Hi!
>>>
>>> On Wed, 14 Jan 2009, Miklos Maroti wrote:
>>>
>>>> Channel 11 is not too bad either, no? Miklos
>>>
>>> Zigbee channel 11 is overlapping with the WiFi channel 1. That is  
>>> no true
>>> for Zigbee channel 26. As you can see from figure 4 from the paper  
>>> the
>>> lossrate of the 4 Zigbee channel overlapping a WiFi channel will  
>>> have losses
>>> if the WiFi traffic is increased. As figure 1 indicate, Zigbee  
>>> channel 24 is
>>> the last one overlapping with the WiFi channel 11. So both 25 and  
>>> 26 should
>>> be the best one, with 26 being the farthest away.
>>
>> Well, maybe the RF230 has some problems with channel 26, so channel  
>> 25
>> could be much better. Anyways, if there are observable real world
>> problems with channel 26 on the IRIS then no academic reasoning is
>> going to change that!!
>
> I agree. But I want to be able to replicate these problems with  
> channel
> 26 on our side.
>
>> We need to have a default where the mote is
>> working at its peek. If people want interoperability, they are fine  
>> to
>> experiment with other channels.
>>
>> So my question, can someone run real application experiments on
>> channel 11, 25, and 26? Loss rate would be interesting to know. (I
>> have no motes)
>
> It will take some time but we'll run some tests and measure the loss  
> rate
> on Zigbee channel 11 with heavy WiFi traffic on WiFi channel 1 and the
> loss rate on channels 25 and 26 with heavy WiFi traffic on WiFi  
> channel
> 11. :-)

Take a look at HyungJune's paper in IPSN 2007. WiFi corrupts 15.4  
packets.

We could switch to 25 as the default if there's a problem with IRIS.  
But 11 is taking a step backward.

Phil
_______________________________________________
Tinyos-help mailing list
[email protected]
https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help

Reply via email to