Li, Chanjuan wrote:
>>> there is another question on tdma configuration. I am not clear the
>>> tdma.conf. I make an example to state my understanding.
>>> I write a simple tdma.conf as follows:
>>> ############################
>>> master:
>>> cycle 500
>>> ip 192.168.1.25 
>>> slot 0 0
>>> slot 1 300 1/2
>>>
>>> slave A:
>>> ip 192.168.1.26
>>> slot 0 100
>>> slot 1 300 2/2
>>> ##############################
>>> I think the config file means time slot like this:
>>> 0 ~ 100 : master for RT traffic
>>> 100~200 : slave A for RT traffic
>>> 200~500 : for master and slave with NRT traffic
>>> ________  ________ _________
>>> |master  | slave A|master   | 
>>> | RT     |  RT    |NRT      |
>>> |        |        |---------| 
>>> | slot 0 | slot 0 |slave NRT|
>>> -----------------------------
>>> 0        100      200
>>>
>>> is it right or not?
>> Yep. As you correctly visualized, the third slot is alternated between 
>> the master and the slave.
>>
>> Note that the actual slot length depend on the maximum packet size for 
>> the slots. As you didn't specify any, TDMA will allow what Ethernet is 
>> able to transmit: 1500 byte payload. In this case, the frame itself can 
>> already take ~122 us (on 100 MBit/s), and you have to account for 
>> transmission jitters. However, if your RT applications on master and 
>> slave do not use the full length, then you can be fine with this config.
> Now, I make a experiment to test performance of rtnet. I just use two
> nodes and do ping-pong to test round-trip time on the top of rtai. the
> result indicat that the round-trip time is almost two tdma cycles no
> matter of packet size,except the packet size around 1500. Is it
> reasonable? I think it can reduce to one cycle if I deploy tdma config
> properly. I use the tdma config as above. Is it possible?

Yes, the can be achieved if the second node as enough time to prepare
the reply between receiving the request and its own next transmission slot.

Jan

Attachment: signature.asc
Description: OpenPGP digital signature

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
RTnet-users mailing list
RTnet-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/rtnet-users

Reply via email to