Community
- Command Central
- MVP Program
- Monthly Mission
- Blogs
- Groups
- Events
- Media Vault
Products
- Observability
- Network Management
- Application Management
- IT Security
- IT Service Management
- System Management
- Database Management
Content Exchange
- SolarWinds Platform
- Server & Application Monitor
- Database Performance Analyzer
- Server Configuration Monitor
- Network Performance Monitor
- Network Configuration Manager
- SQL Sentry
- Web Help Desk
Free Tools & Trials
Store

Latency in ICMP Path Jitter operation

rsdman

Hello

I am trying to understand statistics that VNQM shows for ICMP Path Jitter operation.

My router returns RTT, Jitter and Packet Loss. In VNQM we have Latency in addition.

Could anybody explain me where it comes from and how it's calculated, please?

What demo shows for one of the operations does not make much sense for me (compare RTT and Latency)

Thanks

Find more posts tagged with

Accepted answers

All comments

Radim.Keseg

Hi rsdman

the calculation of latency is done following way, it is a bit hard to describe it precisely so let me put a short form. The values are parsed from CLI command output this way

from CLI: parsing

show ip sla statistics xxxxx

Hop IP ip

RTT : rtt

Number of (RTT|Completion Time): numrtt

Number of Echos - numEchos

PacketLoss: packetLoss

OutOfSequence: outOfSequence

DiscardedSamples: discardedSamples

(RTT|Completion Time) Sum/Sum2: sumRtt/sum2Rtt

if numrtt does not exist then directly by parsing the output then

numrtt = {numEchos - packetLoss - outOfSequence - discardedSamples)

LATENCY = sumRtt/numrtt

RTT = rtt

the LATENCY being longer than RTT might be caused by the fact the RTT is basically actual value measured by the IPSLA operation, however the latency is calculated average of several latest RTTs

Regards, Radim.

rsdman

Hi Radim,

Thank you for providing valuable information!

However, your explanation of Latency calculation is very arguable.

Consider output from the router (lets take one hop as example):

Hop IP 10.2.0.1:

Round Trip Time milliseconds:

Latest RTT: 216 ms

Number of RTT: 12

RTT Min/Avg/Max: 164/216/276 ms

RTT Sum/Sum2: 2596/578448

Jitter time milliseconds:

... Skipped for short

Packet Values:

Packet Loss (Timeouts): 0

Out of Sequence: 0

Discarded Samples: 0

The Last RTT value is calculated here exactly as Latency in your explanation: RTT Sum/Number of RTT = 2596/12=216

So, two questions:

1. For what purposes there is another Latency metric that is calculated exactly (considering your formulas) as RTT?

2. Why this metric shows different, inconsistent values with RTT while should show the same?

Thank you very much,

Alex.

Radim.Keseg

Good question Alex, thanks for asking,

There are certain differences in revisions of device IOS, where not always all results are accessible since the ipsla and rtr syntax differs, therefore the calculation is done as described.

Generally RTR syntax does not contain number of RTT and IPSLA is missing number of echos, in this cases we need to find the value by different approach.

In your case the numbers are showing: the latest RTT is 216ms and it is matching the RTT average in the case, it is relatively ideal case, the latest RTT value might have been in range from 164 to 276, then the latency would be giving different result either.

The values in your screenshot are coming from our Solarwinds demo page, which might show extremely unstable path at given moment, in case you look at different operations via different regions the values are much closer, which reflects more stable environment

Regards, Radim.