[Tinyos-help] CTP: ETX values vs. LPL wake-up interval

Manjunath Doddavenkatappa doddaven at comp.nus.edu.sg
Fri Nov 19 23:23:27 PST 2010


Dear Sir,

   Please find the file at the following link. The file contains PRR vs. 
LPL_wakeup_interval observations. Please skim through the details provided 
below each of the tables in the file.

 	http://www.comp.nus.edu.sg/~doddaven/lpl.pdf

Sincerely,
Manjunath D

################################################################################################################
***************************************************************************************************************

On Wed, 17 Nov 2010, Philip Levis wrote:

> I don't think this quite answers my question: it was whether, on using setLocalWakeupInterval(), you see a change in PRR on simple AM transmissions. I.e., is this a general link layer issue, or an interaction between the link layer and CTP? My guess is that it is the former, but evidence would help.
>
> Phil
>
> On Nov 16, 2010, at 9:55 PM, Manjunath Doddavenkatappa wrote:
>
>>
>> Dear Sir,
>>
>>  The problem and its cause can be summarized as follows.
>>
>>  In the case of CTP, I guess it is required to set LPL variables (local and remote sleep intervals) from the Makefile. Otherwise, I guess no preamble is sent before transmitting routing beacons thus resulting in unexpected behavior. Instead of defining LPL variables in the Makefile, I was using LowPowerListeniing interface to set LPL variables from the application. The interface allows to set local sleep interval, and remote wake-up interval for only the data packets that I send from the application and this does not have any effect on beacon transmission.
>>
>>  As you suggested, I tested without CTP, program works without any problems in both the cases of on setting LPL variables from the Makefile and application. As no routing beacons are involved, behavior is independent of the location from where I set LPL variables.
>>
>> Please correct me if my understanding is wrong.
>>
>> Further, what is the reason for not having the LPL to use default preamble length for outgoing packets based on local sleep interval (that is set using the command setLocalWakeupInterval()). I understand that the command setRemoteWakeupInterval allows to handle different sleep intervals. But is is typical to have different schedules for different nodes ? and moreover, this requires every forwarder to know the sleep interval of its next hop.
>>
>> Thanking you,
>> D. Manjunath
>>
>> ################################################################################################################
>> ***************************************************************************************************************
>>
>> On Mon, 15 Nov 2010, Philip Levis wrote:
>>
>>>
>>> On Nov 7, 2010, at 3:40 AM, Manjunath Doddavenkatappa wrote:
>>>
>>>>
>>>> Dear Dr. Gnawali,
>>>>
>>>> Sorry to get back to you so late, we had a paper deadline.
>>>>
>>>> Everything works fine if I set LPL variables from the Makefile as in TestNetworkLpl. Problems arise only on trying to set LPL variables from the program using "setLocalWakeupInterval()". Just to verify in the later case, I retrived sleep interval value using "getLocalWakeupInterval()", the returned value is consistent with what I set it to.
>>>>
>>>> Manjunath D
>>>
>>> Here's a simpler question: outside of CTP, is there a relationship between the LPL interval and a link's packet reception ratio? I.e., if you send unicast LPL messages to a destination, do you see that increasing the LPL interval decreases the PRR? Note that CTP is often going to be using borderline links that are near the SNR/PRR threshold.
>>>
>>> Phil
>>>
>>>
>
>


More information about the Tinyos-help mailing list