Installation guide

84 Estimating Network Performance
890 USE 100 00
3.13 Predicting MSTR Response Time
When you have calculated the average token rotation time on the
network, you can predict the average time for a response to an MSTR
data request. The response time will not include factors such as
queueing or error conditions on the network. The time will be based on
a requestresponse transaction on a single network. The average
response time is the sum of the following times:
V 1 token rotation time
V 1 scan time of the requesting unit
V 1/2 scan time of the responding unit
The worst case response time would be:
V 2 token rotation times
V 2 scan times of the requesting unit
V 1 scan time of the responding unit
If the scan time of the responding unit is much shorter than the
network’s token rotation time, it is possible for the unit to create the
data response and have it ready in the peer processor before the
network token arrives at the unit. In this case the response would be
transmitted when the token is received. The scan time of the
responding unit can be removed from the timing calculation.
Figure 36 shows an example of a network of six nodes, with the
planned loading. In this example, nodes 1 ... 4 will transmit using
MSTR functions and global data as shown in the figure. Nodes 5 and 6
will not send data in this application but will use the global data when
they receive it.