What a great program! Just got it and have already diagnosed a few nagging problems with sites that have been experiancing lots of choppinessand drops in voice quality. The drops were easy to spot as the red lines showed the incidents clearly. Two questions;
1) How can I show jitter through the Web feature?
2) With VoIP the most common issue I run into - after latency and flakey routers - is the issue of having enough bandwidth up and down (usually up). What would you recommend to show the customer definitively that he's losing packets - especially voice packets - as a result of having too skinny a connection? Would PPPro be enough or would you recommend something to suppliment it?

Thanks,

Eric K

Hi, Eric.

1) The web interface should have a column for the jitter. If it doesn't show up, right-click on the column header and you should have an option to turn it on. Of course you need to have the Jitter script enabled (which is probably the case already - it sounds like you have it showing in the GUI already).

2) Inadequate bandwidth can happen at so many points and so many places - it's hard to tool-up all the points needed to measure bandwidth shortcomings at every place where it could happen.

PingPlotter doesn't jump up and down and wave red flags at every bandwidth shortcoming in the route, but some experience with your route's data (plus PingPlotter graphs, of course) is extremely effective at locating bandwidth problems.

You may have already read our voice over IP troubleshooting guide, but that guide spends most of its time talking about how to identify bandwidth limitations (congestion) using PingPlotter graphs - especially when using time-of-day as an indicator of problems (if performance varies based on time-of-day, that's a pretty strong indicator of bandwidth problems).

If you're looking for a smoking gun, one way to do this is to hook up to the device that you suspect is constraining bandwidth and measure traffic flow on that device. If the traffic rate is hitting the subscribed bandwidth limit, then that's a smoking gun when paired with a PingPlotter latency graph. The problem with this is that it's not generically available with all consumer devices - each one has different ways of exposing bandwidth use (or no way at all).

If you have a device that is pollable with SNMP, get in touch with us for a script that can be plugged in to PingPlotter Pro to query bandwidth use via SNMP and put it on a graph in PingPlotter Pro.

- Pete

Pete -
Another VoIP question. I remember reading some where in the PPP3 literature that I could send a payload so that it looked like a voice packet. Where does it talk about setting that up? Great program btw.

Eric K

Hi, Eric.

There are a couple of places for this.

First, some VoIP uses a ToS setting in the DSCP field of the packet. This can be configured in the Engine Settings, ToS / DSCP Header byte. You'll want to look at your VoIP stream to see what's being used there.

Another thing you can do is to capture a VoIP packet and use that for the Cargo of your packet. This option will fill the packet with data of your choosing to more closely simulate the application you're using. This is configured in the same area (Engine Settings), in "Advanced cargo options".

We don't currently have any specific recommendations on values you can use here. You can use something like WireShark (used to be Ethereal) to capture your packets, and then use the captured data to build a Cargo and DSCP value that makes sense for your application packets.

- Pete