Visit our LabVIEW Wiki Real Time article
Tags |
(This content has not been tagged yet)
|
![]() |
Aug 5 2004, 12:57 PM
Post
#1
|
|||
|
More Active Member Posts: 41 Joined: 11-May 04 From: Italy Member No.: 399 Using LabVIEW Since:2000 LV:4.0 ,7.1.1 ,8.20
|
I need to develop a network distributed application for data monitoring and control so i must create the comunication protocol.
I think to implemet my protocol over UDP protocol using standard ETH but i need a fast and deterministic comunication, do you have experience in this application of ethernet? Bye, Alessandro. PS sorry for my english --------------------
|
||
|
|
|||
| Ad |
Aug 5 2004, 12:57 PM
Post
#
|
||
|
|
|
||
|
|
|||
Aug 7 2004, 10:32 PM
Post
#2
|
|||
|
Very Active Member Posts: 164 Joined: 18-March 04 From: Canada Member No.: 311 Using LabVIEW Since:1995 LV:8.20 ,7.1.1 ,.
|
QUOTE (Ale914 @ Aug 5 2004, 07:57 AM) I need to develop a network distributed application for data monitoring and control so i must create the comunication protocol. Hi Alessandro, This is not a direct answer to your question... just want to make you aware of an existing product -- the Datalogging and Supervisory Control (DSC) module for LabVIEW, which has built-in networking functionality for monitoring and control. Please visit the following page for more info' on the LabVIEW-DSC module: http://sine.ni.com/apps/we/nioc.vp?cid=10418&lang=US Regards, Khalid Ansari
|
||
|
|
|||
Aug 12 2004, 08:35 AM
Post
#3
|
|||
|
Extremely Active Member Posts: 391 Joined: 6-February 04 From: switzerland Member No.: 253 Using LabVIEW Since:1997 LV:7.1.1 ,6.1 ,.
|
Alessandro,
UDP (as well as TCP) is a non-deterministic protocol. But, many tests were made to still use TCP/UDP as a deterministic system: - If you have just 2 partner (with full duplex crossed ethernet cable) then deterministic is no problem, because TX of on computer is connected to RX of the other. You will never see collisions. - If you have more than 2 partner, build local segments in your ethernet the way that in a segment the traffic never rise over about 5%. Then collisions are so rare that you almost can forget them. Another tip: If you just collect data, - either use Khalid's proposal or - use for the data acquisition stations intelligent systems (such as Fieldpoint/PXI with LabVIEW real-time), set on this intelligent system a timestamp for every datapoint and then send it over the ethernet. This way your central station always knows when something occured and it can link the data of the different stations in correct order (even if the data arrives 1-2 seconds later). Didier
|
||
|
|
|||
![]() ![]() |
| Time is now: 22nd November 2008 - 07:09 AM |