Is there some publications or statistics to justify the use of small
packets  ? (especially for micaz and telosb)

2009/7/28 Raffaele Gravina <[email protected]>

> Longer messages will reduce the data overhead introduced by headers.
> However, shorter messages have a smaller error probability and the cost of a
> retransmission would be lower. Chosing the largest message size is not the
> best solution ever. For instance, higher messages could result in longer
> delays.
>
> I read somewhere that 30-31 bytes would be a good trade-off.
>
> Using telosb and TinyOS 2.1, the largest TASH_DATA_LENGTH i was able to set
> is 114 bytes.
>
> Cheers,
>
> Raf
>
> 2009/7/28 Rémi Villé <[email protected]>
>
>> Hi,
>>
>> TOSH_DATA_LENGTH is about 20-30 byte, depending of the mote. I have read
>> that we can increase this value to something like 100 byte.
>> But what would be the reason to choose a 100 byte payload rather than a 30
>> byte one ?
>> I mean if I have 100 byte to send, should I send 4 packet of 30 byte ?
>> Maybe because if there's one error in the message (CRC detection) I should
>> retransmit 100 byte rather than 30 byte...
>>
>> I don't know which is the best way to choose a good TASH_DATA_LENGTH
>> (assuming that I have an infinite number of byte to send), maybe by
>> experimentation or by analyzing? or maybe the default length has been chosen
>> with this considerations?
>>
>> Thanks,
>>
>> Rémi
>>
>> _______________________________________________
>> Tinyos-help mailing list
>> [email protected]
>> https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help
>>
>
>
>
> --
> Raffaele Gravina | Research Engineer
>
> Wireless Sensor Networks Lab Berkeley
> 2000 Hearst Ave, suite 304
> Berkeley, CA 94709
> +1 510 666 0174 ext. 101
>
_______________________________________________
Tinyos-help mailing list
[email protected]
https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help

Reply via email to