Application Response Time Examples

Introduction

The following examples are designed to help you calculate the application response time (ART) for an application.

Example 1: A 140CRA31200 in a Remote I/O Main Ring

This example calculates the ART from the perspective of the 140CRA31200 adapter module in a remote I/O sub-ring in the following Quantum EIO network design:

primary local rack

secondary Hot Standby rack

remote I/O drop daisy-chained on the remote I/O main ring

In this example, ART is calculated from the perspective of the adapter in the one of the remote I/O main ring drops (device 4, above). Consider the following application-specific elements when calculating ART:

The maximum potential hop count — i.e., the maximum number of switches a packet might need to pass through from the remote I/O adapter (4) to the remote I/O head module in the local rack (1) — is 4. This would be the case if a packet follows the path from the remote I/O drop (4), through another remote I/O drop (3), through the secondary rack (2), to the primary local rack (1).

NOTE: The hop count includes all switches located along the route between the source input module and the CPU, including the switches embedded in the CRA and CRP modules.
Jitter is introduced into the system only from the 2 remote I/O main ring drops.

Given these factors, ART computation parameters include:

Parameter	Maximum value (ms)	Comments
CRA drop process time (CRA_Drop_Process)	4.4	The sum of CRA input scan time and queue delay.
CRA input RPI (RPI)	–	User defined. Default = 0.5 * CPU period.
network input time (Network_In_Time)	(0.078 * 4) = 0.312	Hop-count is 4 from the CRA in remote I/O drop (4) to the CRP in local rack (1), which includes the switches in both the CRA and the CRP.
network input jitter (Network_In_Jitter)	(0.078 * 2) = 0.156	For delay occasioned by devices (2) and (3).
CRP input jitter (CRP_In_Jitter)	0.6 + (3 * 0.2) = 1.2	To read packet and forward it to the CPU.
CPU scan time (CPU_Scan)	–	User defined, based on application.
CRP output jitter (CRP_Out_Jitter)	1.6	CRP internal queue delay.
network output time (Network_Out_Time)	(0.078 * 4) = 0.312	Hop-count is 4 from the CRA in remote I/O drop (4) to the CRP in local rack (1), which includes the switches in both the CRA and the CRP.
network output jitter (Network_Out_Jitter)	0	Does not apply. No distributed I/O devices are connected to the remote I/O network.
CRA drop process time (CRA_Drop_Process)	4.4	The sum of CRA output scan time and queue delay.
For an explanation of each parameter, refer to the ART Computation Parameters topic.

Recall that the ART formula is:

ART= (2*CRA_Drop_Process) + (RPI) + (Network_In_Time) + (Network_In_Jitter) + (CRP_In_Jitter) + (2*CPU_Scan) + (Network_Out_Time) + Network_Out_Jitter)

Thus, for a CPU scan time of 50 ms and RPI of 25 ms, maximum ART is:

max ART = (2*4.4) + 25 + (2*0.312) + .156 + 1.2 + (2*50) + 1.6 = 137.380 ms

Example 2: A 140CRA31200 in a Remote I/O Sub-ring

This example calculates the maximum ART, representing the longest packet path from a 140CRA31200 adapter module in a remote I/O sub-ring, to the remote I/O head module in the local rack. The calculation is performed from the perspective of the 140CRA31200 adapter module (9) in the following Quantum EIO network design:

primary local rack

secondary Hot Standby rack

DRS on the remote I/O main ring

distributed I/O sub-ring with 2 distributed I/O drops

remote I/O drop daisy-chained on the remote I/O main ring

HMI device off the remote I/O drop

distributed I/O device (cloud)

remote I/O drop with 140CRA31200 adapter module

In this example, ART is calculated from the perspective of the adapter in the remote I/O drop (device 9, above). Consider the following application-specific elements when calculating ART:

The maximum potential hop count — i.e., the maximum number of switches a packet might need to pass through from the remote I/O adapter (9) to the remote I/O head module in the local rack (1) — is 6. This would be the case if a packet follows the path from the remote I/O drop (9), through a DRS (4), through the CRA in a remote I/O drop (6), through a second DRS (3), through the secondary rack (2), to the primary local rack (1).

NOTE: The hop count includes all switches located along the route between the source input module and the CPU, including the switches embedded in the CRA and CRP modules.
Jitter, also known as packet queue delay, is introduced into the system by the following design elements:
- the distributed I/O sub-ring (5)
- the remote I/O sub-ring, on which the remote I/O adapter (9) is located
- the remote I/O drop (6)
- the HMI (7)
- the distributed I/O device (8)

Given these factors, ART computation parameters include:

Parameter	Maximum value (ms)	Comments
CRA drop process time (CRA_Drop_Process)	4.4	The sum of CRA input scan time and queue delay.
CRA input RPI (RPI)	–	User defined. Default = 0.5 * CPU period.
network input time (Network_In_Time)	(0.078 * 6) = 0.39	Hop-count is 6 from the CRA in remote I/O drop (9) to the CRP in local rack (1), which includes the switches in both the CRA and the CRP.
network input jitter (Network_In_Jitter)	((0.078 * 2) + (0.128 * 4)) = 0.668	2 remote I/O packets from devices (6) and (2) plus 3 distributed I/O packets from distributed I/O devices at locations (5), (8) and (1)
CRP input jitter (CRP_In_Jitter)	(0.6 + (2 * 0.2)) = 1.2	To read and forward packets from devices (9) and (6).
CPU scan time (CPU_Scan)	–	User defined, based on application.
CRP output jitter (CRP_Out_Jitter)	1.6	CRP internal queue delay.
network output time (Network_Out_Time)	(0.078 * 6) = 0.39	Hop-count is 6 from the CRA in remote I/O drop (9) to the CRP in local rack (1), which includes the switches in both the CRA and the CRP.
network output jitter (Network_Out_Jitter)	(0.128 * 4) = 0.512	Worst case, for devices (4), (6), (3), (2), and (1).
CRA drop process time (CRA_Drop_Process)	4.4	The sum of CRA output scan time and queue delay.
For an explanation of each parameter, refer to the topic ART Computation Parameters.

Recall that the ART formula is:

ART= (2*CRA_Drop_Process) + (RPI) + (Network_In_Time) + (Network_In_Jitter) + (CRP_In_Jitter) + (2*CPU_Scan) + (Network_Out_Time) + Network_Out_Jitter)

Thus, for a CPU scan time of 50 ms and RPI of 25 ms, maximum ART is:

max ART = (2*4.4) + 25 + (2*0.39) + .668 + 1.2 + (2*50) + 1.6 + .512 = 138.560 ms

NOTE: If a cable break exists on the network, add an additional time period — equal to 50 ms + RPI — to the above ART calculation. The added time allows for RSTP recovery from the cable break.