Can other users post their latest TBB Quality Monitor graph please?

[Bill]

EDIT - Looking again the period immediately after midnight is definitely heavy downloading as the minimum latency also increases massively. The rest of the day definitely shows sign of persistent internet use.

Do you mean "packet loss" rather than "minimum latency"? That's the off-peak period cutting in. Interesting that Steve doesn't seem to get it on his trace

Mine's behaving nicely now:

[JamesAllen]

That looks like typical P2P downloading over a long period.

Do you have more than one PC? Are there other family members with access to a PC? Do you use wireless? If so how confident are you it's secure?

EDIT - Looking again the period immediately after midnight is definitely heavy downloading as the minimum latency also increases massively. The rest of the day definitely shows sign of persistent internet use.

Hi there,

I work from home so I'm using the connection constantly all day (often including internet radio streaming) . I do also have background downloads, though in the day (prime time) it's only using upstream and usually not at super high throughput - though I assume that can still affect latency?

Am I right to assume that most of the graphs here showing low latency are connections that do not have large levels of constant use then?

The minimum latency spikes on the graph yesterday were to do with some downloads beginning at midnight, but I was also gaming at that time as well.

Just trying to get my head around exactly what I'm looking at on my graph.

[esh]

Asymmetric connections will always be heavily affected by upload in my experience, usually causing a very large increase in ping jitter (so the average will go up, though technically the minimum ping should be the same). I have found that when downloading at 6Mbit, you need more than 25% of the 830k upload you get to sustain the traffic.

[JamesAllen]

Asymmetric connections will always be heavily affected by upload in my experience, usually causing a very large increase in ping jitter (so the average will go up, though technically the minimum ping should be the same). I have found that when downloading at 6Mbit, you need more than 25% of the 830k upload you get to sustain the traffic.

Ah certainly. I have always throttled uploads etc to avoid it killing the downstream throughput.
So should I not be seeing the minimum ping varying like this? As my upload is far from being maxed out but I'm seeing such a huge jumps in minimum ping. I suppose I just want to confirm that what I'm seeing is normal for a connection that is being used but without upstream being saturated.

[esh]

The last graph I saw of yours had a fairly static block of green indicating a stable minimum ping but a wildly varying maximum ping. That is pretty much what I would expect for an active connection. Just so you can compare I attach here two of my quality charts indicating an "idle" period of my connection and a "moderate" in-use period with my connection.





The minimum ping changed by 1.5ms, but I wager that is more to do with time of day than use. The average has increased by ~5ms and you can see the jitter is much higher (there is also one packet loss blip, spot on the hour).

[JamesAllen]

Aha, thanks esh - that helps a lot. Good to know my connection is performing as expected. Certainly had no issues streaming internet radio since the router reboot.

[esh]

Just for the record, I've noticed I'm getting very close to my 7.1Mbit back again now, not sure if the idnet tweaking did any help there.

[faircot]

What's happened here to cause the drop in latency at 10pm last night, IDNet or BT? I'm not complaining; any drop is welcome. My graph has been constant ever since I started it months ago.
Just curious.

[Bill]

What's happened here to cause the drop in latency at 10pm last night, IDNet or BT?

I've had that happen (in both directions!) occasionally. It indicates a change in routing somewhere along the line, but I'd say your chances of finding out who did it are close to zero

It'll probably go back up again at some point in the future (could be hours, could be weeks), mine always did

[faircot]

It'll probably go back up again at some point in the future

A bit like my overdraft then

[Rik]

 

[esh]

I suppose the routing change could be from TBB's point of view as well.

[wdforte]

May I butt in here? I am still trying to set up this BB Quality Monitor thingy via TBB and am asked for an IPv4 or IPv6 address to complete the task. I have no idea what these are and even less on how to establish which one I am on, if any?  Can anyone help please? 

[Ray]

You'll need the IPV4 address, and it's the IP address provided to you by Idnet.

[Simon]

Your IPv4 address should show in the bottom right hand corner  of your posts, if that helps.  Starts with 212.

[wdforte]

Thanks Chaps Thought that was (212.69******) simply my IP address Thanks again.

[JamesAllen]

Hi guys,

Just wanted your take on my graph. Now I was streaming internet radio - two 128kbs stream from 2pm today until 9pm. However, during the broadcast in the evening I had some issues being able to send fast enough to one of the encoders. Not sure if was the destination server or on my side.

Does this graph look ok for upstream at about 30-40 k/sec?

The average latency looks abit odd to me but I'm still not fully used to what I should be seeing under various circumstances.

[Rik]

What were you doing before 2pm?

[JamesAllen]

What were you doing before 2pm?

Just using my connection normally really. There may have been some upstream being used though to be fair.

[Rik]

The thing is, James, pings mean nothing in isolation. If you really want to find out how your line is behaving, turn the computers off for 24 hours. That graph will tell you about line performance, anything else tells you nothing unless you also log your WAN usage at the time.

[JamesAllen]

The thing is, James, pings mean nothing in isolation. If you really want to find out how your line is behaving, turn the computers off for 24 hours. That graph will tell you about line performance, anything else tells you nothing unless you also log your WAN usage at the time.

Thanks Rik. But from the graph I posted, that looks as expected for using less than half of my upstream would you say? Just trying to work out if perhaps it was my connection that caused my streaming issue during broadcast last night.

[Rik]

I can't say what it looks like, James. The only graph that gives a picture of your line's performance is a 'no use' graph. Everything else reflects what you were doing at the time, to a greater or lesser degree, so without knowing the exact load you had on the connection, all anyone can do is speculate.

[esh]

The 5pm-9pm bit looks awful, and is about the time congestion would kick in too, plus there is packet loss.

Uploading about 10KB/s on an 832kbit up connection here, my average ping doubles from 15ms to 30ms, 40ms at the outside. Mid-morning your average latency is 40ms but goes up to 140ms at 7pm. I think most telling is the spikes in *minimum* latency around this mark too. However, I would wager it is the packet loss that would mostly mess up streams. If you only have the 384 up line, then 2x128 kbit streams is going to be pushing it. Should be fine on an 832 I would think though, as long as you aren't on a full speed download at the time. To be honest, I would say it looks like more than just a moderate upload, either because you have PCs doing other things or the exchange is full up. Given the times when it is worst, I would say exchange probably comes into it.

[JamesAllen]

The 5pm-9pm bit looks awful, and is about the time congestion would kick in too, plus there is packet loss.

Uploading about 10KB/s on an 832kbit up connection here, my average ping doubles from 15ms to 30ms, 40ms at the outside. Mid-morning your average latency is 40ms but goes up to 140ms at 7pm. I think most telling is the spikes in *minimum* latency around this mark too. However, I would wager it is the packet loss that would mostly mess up streams. If you only have the 384 up line, then 2x128 kbit streams is going to be pushing it. Should be fine on an 832 I would think though, as long as you aren't on a full speed download at the time. To be honest, I would say it looks like more than just a moderate upload, either because you have PCs doing other things or the exchange is full up. Given the times when it is worst, I would say exchange probably comes into it.

Thanks esh - that is very helpful indeed. I have 1136 upstream so more than enough to handle two concurrent 128k stream luckily, although to be fair I was also doing a ustream webcam stream as well which is only going to push the latency up further I would imagine.

It's just that I've been broadcasting during this slot for months now and this is the first time I had a stream failure, so wanted to look into whether it could be my connection or the remote server.

More so because of the recent problems we have all been having. It's a total nightmare in a live show to lose connectivity and completely ruins the flow of the broadcast. Hopefully it won't happen next week though or I can identify any issues that may be present.

@Rik Understand what your saying but it's either post what I have or not bother at all, which means I then get no insight into what might be happening. It's all a learning experience..

[Rik]

No-one can give you a definitive answer as long as you're loading the connection, James. If you talk to support, they'll either ask you to do what I'm suggesting or to boot into safe mode with networking support. Without knowing what processes are using the 'net on your machine, all anyone can do is give you a general answer.