Kitz Forum
Broadband Related => Router Monitoring Software => Topic started by: roseway on August 12, 2012, 04:07:28 PM
-
Linux 32-bit and 64-bit versions, Windows 32-bit version.
Changes compared to v4:
- corrected errors in autosaving snapshots
- added option to display bitloading chart in expanded or compressed style
- changed filename of config file in Windows version to 'rsw.ini'
- minor cosmetic changes
https://docs.google.com/folder/d/0BxbUtOYVZ_SCZ1BnY3RKYnN2b1U/edit
(I've removed some of the earlier versions, although they're still available in the unlikely event anyone wants them).
-
Thank you Eric. I should be able to test the 64-bit proper version straight away. ;)
-
Any plans for a Mac version?
-
Any plans for a Mac version?
Unlikely, I'm afraid, unless someone could provide me with a Mac and a suitably qualified programming assistant. :)
-
I am happy to say that the lake snapsnot problem has now been resolved. Also the regular & expanded Bitloading graphs are produced, as expected.
I have deliberately configured the directory location, to which snapshots and periodic graphs are saved, without the suffix "/". I shall have an answer in less than an hour's time.
-
The option to view bit-loading in compressed & expanded views is really useful, particularly for VDSL2 connections that use up to 16 times as many tones as ADSL1 & ADSL2 connections.
I have set up & am currently awaiting the automated snapshots, so I have attached manual bit-loading snapshots to demonstrate the compressed & expanded views accordingly.
The blue "shared or other" bit loading now looks correct too.
-
I am happy to report that I have now had a successful trio of graphs, generated automagically. :)
So it's :thumbs: from The Cattery.
-
Thanks for confirming it. :)
-
I can also confirm that auto-generated snapshots are now working on my Windows 7 machine.
I am using the default 2 hour window, but snapshots were generated at these times (not quite 2 hour intervals):-
21:21
23:16
01:11
03:06
05:01
06:56
08:51
10:46
The auto generated Bit-loading snapshot alternates between compressed or expanded view (whichever is selected in the GUI - or not).
It might be nice to chose which to default to, regardless of which version was last viewed in the GUI.
For my VDSL2 connection, I hoped to generate lots of compressed view for analysis/comparison, but I forgot to switch back to compressed view & thus now have a mixed bag of snapshots.
I can always start again & rely on my memory to force the correct version for auto generated snapshots, but................. ;)
-
I can also confirm that auto-generated snapshots are now working on my Windows 7 machine.
Thanks for confirming that.
I am using the default 2 hour window, but snapshots were generated at these times (not quite 2 hour intervals):-
21:21
23:16
01:11
03:06
05:01
06:56
08:51
10:46
I allowed 5 minutes overlap between snapshots to make sure that nothing was lost in rounding errors and processing time, but it's probably a lot more than is necessary.
The auto generated Bit-loading snapshot alternates between compressed or expanded view (whichever is selected in the GUI - or not).
It might be nice to chose which to default to, regardless of which version was last viewed in the GUI.
That's a good point. I'll look at it.
I'm going to try to improve the presentation of the compressed snapshots. The one which you showed earlier had a lot of wasted space at the right hand side where there was no bitloading. I plan to trim the extent of the chart to just the range of tones actually used (plus a few).
And I've had a closer look at the profile bands, and I think that I've got it a bit wrong still. I've looked at the band info which you provided, and also the Digital Region bands, and I find that the only area of overlap is between tones 32 and 95. In the higher bands, there are upstream/downstream crossover points at tones 868, 1208, 1970 and 2780 which accommodate all the band plans with no overlap.
-
I'm going to try to improve the presentation of the compressed snapshots. The one which you showed earlier had a lot of wasted space at the right hand side where there was no bitloading. I plan to trim the extent of the chart to just the range of tones actually used (plus a few).
It's only wasted on my connection as attenuation over distance from the cabinet is too high to actually use the higher tone bands.
See the attached for a connection that CAN actually use them.
Just look at the Attainable rates.
That connection is only around 80m from the cabinet though.
It's also useful to see what is/isn't being used from the theoretical maximum, so personally, I would show all 4096 tones along the x axis, whether actually used or not.
It really highlights the difference between good & poor connections.
And I've had a closer look at the profile bands, and I think that I've got it a bit wrong still. I've looked at the band info which you provided, and also the Digital Region bands, and I find that the only area of overlap is between tones 32 and 95. In the higher bands, there are upstream/downstream crossover points at tones 868, 1208, 1970 and 2780 which accommodate all the band plans with no overlap.
Yes, but there are also some "other" populated (not overlapped) tones between "official" band plans (i.e. between DS tone 859 & US tone 868 on a Huawei DSLAM).
Cheers,
Paul.
-
It's also useful to see what is/isn't being used from the theoretical maximum, so personally, I would show all 4096 tones along the x axis, whether actually used or not.
It really highlights the difference between good & poor connections.
OK, I'll leave it as it is then.
Yes, but there are also some "other" populated (not overlapped) tones between "official" band plans (i.e. between DS tone 859 & US tone 868 on a Huawei DSLAM).
Oh I see, I didn't realise that some of the unallocated tones are actually used. They'll have to be blue then.
-
Eric I may be doing something wrong or not understanding but I see no option to set the snapshot interval, unless it uses the time per page setting. Also when started with no existing settings the snapshot default folder still has a trailing slash. This is the 32 bit Linux version.
Stuart
-
Stuart,
The snapshot interval is the time per page (minus 5 minutes at present, but will be minus 2 minutes). The snapshot directory needs a trailing slash, so that's what it has by default; however, because the user can enter a path manually, possibly without the trailing slash, the program can cope with either situation. It's not causing a problem is it?
-
Stuart,
The snapshot interval is the time per page (minus 5 minutes at present, but will be minus 2 minutes). The snapshot directory needs a trailing slash, so that's what it has by default; however, because the user can enter a path manually, possibly without the trailing slash, the program can cope with either situation. It's not causing a problem is it?
No problems at all just observations.... thanks.
Stuart
-
Works beautifully on Ubuntu. Are there any plans for error (e.g. error seconds, CRC, FECs, etc.) monitoring?
-
Thanks for that. Yes, There are plans to monitor other data, including errors.
-
I am having one issue. The program seems to get stuck on one tab and I can't switch to any others unless I restart the program. It happens quite randomly but it seems to be more likely to occur if the program is kept running for a long time.
-
It does become a bit unresponsive while it's collecting a sample from the router. There's a red 'Sampling' indicator at the top, which is displayed while this is happening. Is it possible that this is what you've experienced? It should become responsive again after a few seconds.
One thing I ought to do is rewrite the part of the program which accesses the router as a separate thread, but I've been putting it off because it's quite difficult.
-
It freezes on one tab even when not sampling.
-
Can you confirm that it actually stays frozen, or does it recover if you wait a while?
-
I can confirm that after leaving it over an hour, it remains frozen.
-
OK, thanks. I've no idea what this can be at the moment.
-
Sorry ar545, one more question: did you have automatic snapshots enabled when the freezes occurred? If so, which graphs were set to autosave, and did you try running the program with automatic snapshots disabled, to see if the freezes still happened?
-
It happens with or without automatic snapshots enabled.
-
Thanks, time to put on thinking cap.
-
Any plans for a Mac version?
Unlikely, I'm afraid, unless someone could provide me with a Mac and a suitably qualified programming assistant. :)
Hi Eric, Bald_Eagle1 pointed me in your direction to shed some light on my woes (http://forum.kitz.co.uk/index.php/topic,11424.15.html). Would you be willing to divulge the source code to see if I can get it running on a Mac? I assume the Linux -> Mac porting shouldn't be too difficult if it's written using X etc?
Else, I could bodge a similar result - was thinking a cron job running a simple bash script to collect the necessary data (what I do currently as I am remote from where the installation is) and append the values to a csv file + a gnuplot script to process the output when required.
I'm chasing info on an unusual occurrence that OpenReach have told me of (they see it at their end) - our line speed drops between midnight and 5am apparently, so something is happening somewhere... Be it an electric fence interfering, or perhaps some farm machinery kicking in causing interference, or even a temperature drop, it would be good to get some visibility.
Would appreciate your thoughts,
Tristan
-
Hi Tristan,
There's no secret about the source code - it's open source, but I haven't formalised the licensing yet. It can almost certainly be built on a Mac, because the development system I use (Lazarus/Freepascal) is a multi-platform system which includes MacOS in its supported list. You can't just take the source code and rebuild it on another platform - you need to install Lazarus and do some necessary configuration, then load rs-ux, modify it as necessary for the platform, then build it from there. If you're familiar with Lazarus it's straightforward, otherwise it would be a bit of a challenge.
The current state of development is that I completed a version 0.6, but didn't release it because I decided to do some major rewriting of the graphing sections of the program. So it's in a state of flux, but v0.6 is usable. If you can make use of its source code, let me know and I'll package it up and make it available.