Think broadband monitor

jb66 · 10-01-2012, 11:25

I sometimes see folk posting these when complaining about performance. When I'm Hammering torrents my graph looks poor but when it's idle it's fantastic. How can this be used as a reliable indication of how good someones Internet is if we can't prove someone on there network is torrenting?

---------- Post added at 11:25 ---------- Previous post was at 11:23 ----------

This is my superhub that I'm testing to see if it's any good on 5ghz

This is my vmng300 downloading 30gig of torrents

kwikbreaks · 10-01-2012, 11:28

You can't prove anything but the poster should know. Running torrents or simply downloading something with http, nntp, ftp etc. will make a complete horlix of the chart.

Here's one of mine when I had a period of http downloading - tell me if you can spot when it was...

If you have some trouble spotting it I'll dig one out where it's more obvious.

===update===
I see you've added some charts now.

Sephiroth · 10-01-2012, 12:59

For my purposes the value of the TBB graph is to see what's happening when I'm doing nothing.

qasdfdsaq · 10-01-2012, 15:01

QoS FTW. Or just moderating your downloading.

Tell me if you can spot where I downloaded 80GB... And no, it wasn't 4am.

djfunkdup · 10-01-2012, 15:55

Quote:

Originally Posted by Sephiroth

For my purposes the value of the TBB graph is to see what's happening when I'm doing nothing.

yip same here man......... this is my graph

kwikbreaks · 10-01-2012, 16:15

Quote:

Originally Posted by qasdfdsaq

QoS FTW.

I just kick any downloads off just before I go to bed so I don't give a stuff what it does to my network. I have got a demo webcam and a weathercam both available on the public web and I have to set some pretty draconian QoS rules for them - especially as I'm now on 10Mbps with very restricted upstream limits.

qasdfdsaq · 10-01-2012, 18:00

Yeah but then your TB Ping graph becomes meaningless

(To be fair I don't use QoS either, but my downloads are barely noticeable on my chart for some reason)

Skie · 10-01-2012, 20:20

I use QoS to cap the upload to 90% and download to 97% for a single machine, and have HTTP and DNS prioritized.

That stops my connection from being unusable when steam ramps up its downloads or P2P manages to get full use of the upload.

Chrysalis · 11-01-2012, 09:20

Seems a silly thread, obviously I know when I am doing something that would mess up my graph and during that period I discount recorded data as meaningful.

jb 66 do you think I did any downloading/uploading on this graph?

or this one perhaps?

Sephiroth · 11-01-2012, 10:33

That's kinda cheating, Chrys. Between May 2011 and Jan 2012 you may have been resegmented or something. Or one's the SH and the other's the Ubee.

kwikbreaks · 11-01-2012, 10:59

Quote:

Originally Posted by qasdfdsaq

Yeah but then your TB Ping graph becomes meaningless

Not to me it doesn't as I know when I've done some downloading.

The effects of downloading on mine are very evident doing it my way and presumably evident to anybody else capable of interpreting the chart. Doing it your way and running throttled back downloads for long periods makes your charts meaningless to everybody including you as have no idea what it may look like if you weren't downloading at all and it's only when your connection is idle that the chart throws any light on just the connection.

I suppose I should admit that I don't deliberately download the way I do for any other reason than it requires least effort.

Chrysalis · 11-01-2012, 11:05

heh well.

both are vmng300

the first is from may 2011 when I had uplifted upstream capacity but was still on old upload configs, so wasnt over subbed. The connection was idle all day very easy to see.

the second is from yesterday. Connection was idle all day not so easy to see. Exception been small spikes caused by samknows 4am 10 am 4pm then hourly 6pm to midnight.

the 2 major events that degraded it were the upstream uplift rollout in june and when 100mbit was switched on in sept.

by idle I mean light usage, web browsing, bit of youtubing that sort of thing. Stuff that isnt enough to affect graph, no uploading at all.

---------- Post added at 11:05 ---------- Previous post was at 10:59 ----------

this one is interesting.

The mass of green and blue was me uploading a usb image to a ftp server, the increased jitter after it took me by surprise tho. Rebooting the vmng300 then reduced the jitter noticebly but not all the way.

kwikbreaks · 11-01-2012, 11:09

Quote:

Originally Posted by Chrysalis

the increased jitter after it took me by surprise tho. Rebooting the vmng300 then reduced the jitter noticebly

This makes no sense at all to me. Did you have the same upstream post the reboot as before it?

Chrysalis · 11-01-2012, 11:11

yes the vmng300 wont ever cycle the upstream channels.

It seems when I did that upload it increased the queue size on the upstream (like the superhub) which increased packet delays. But you can clearly see right after the faint red line (Reboot) jitter dropped sharply.

Sirius · 11-01-2012, 12:56

Shub in modem mode with a ClearOS router hanging off it.

://