spdMerlin spdMerlin - v4.x - speedtest and bandwidth monitoring

[heysoundude]

is v4.0.1 pushed yet?

 

[NGI]

@NGI please try updating and try again

Thanks @Jack Yaz it works perfect now.

 

[Jack Yaz]

Does AutoBW only update if you SSH and select 'Update QoS bandwidth value now'?

either that, or if you have enabled option 4 (or 5 if on develop) it applies after every scheduled speedtest

 

[Jack Yaz]

is v4.0.1 pushed yet?

it's now 4.1.0 as a threshold feature for autobw crept in
hoping to release tonight

 

[Jack Yaz]

Thanks @Jack Yaz it works perfect now.

great!

 

[Jack Yaz]

v4.1.0 is now available
Changelog:

• NEW: Set a threshold that must be broken for AutoBW calculated values to be applied over existing values
• FIXED: Data usage over 1GB now reports correctly
• FIXED: Last 10 results table would incorrectly populate columns when result URL collection is disabled
• CHANGED: Use 7za instead of 7z (for older Entware installs)

 

[QuikSilver]

v4.1.0 is now available
Changelog:

• NEW: Set a threshold that must be broken for AutoBW calculated values to be applied over existing values
• FIXED: Data usage over 1GB now reports correctly
• FIXED: Last 10 results table would incorrectly populate columns when result URL collection is disabled
• CHANGED: Use 7za instead of 7z (for older Entware installs)

@Jack Yaz would it be possible to implement a decimal into the MBPS for those who have low connections. I believe it may benefit me for my upload to keep between 1.7 and .8 or so. Thanks!

 

[Jack Yaz]

@Jack Yaz would it be possible to implement a decimal into the MBPS for those who have low connections. I believe it may benefit me for my upload to keep between 1.7 and .8 or so. Thanks!

I'll add it to my list to look into when I'm next working on spdMerlin

 

[QuikSilver]

I'll add it to my list to look into when I'm next working on spdMerlin

Thank you!

 

[NGI]

@Jack Yaz on v4.1.0 when I use "Update QoS bandwidth values now" from the CLI I can currently see a strange 2.4% difference between the values of the DB/UB speed from the CLI compared to the GUI. E.g. Download Bandwidth from CLI "...existing value (24677 Kbps)" from the GUI - 24.0986328125 Mb/s and for the Upload Bandwidth from the CLI "...existing value (5289 Kbps)" and from the GUI - 5.1650390625 Mb/s.

 

[Jack Yaz]

@Jack Yaz on v4.1.0 when I use "Update QoS bandwidth values now" from the CLI I can currently see a strange 2.4% difference between the values of the DB/UB speed from the CLI compared to the GUI. E.g. Download Bandwidth from CLI "...existing value (24677 Kbps)" from the GUI - 24.0986328125 Mb/s and for the Upload Bandwidth from the CLI "...existing value (5289 Kbps)" and from the GUI - 5.1650390625 Mb/s.

by GUI do you mean the adaptive qos page?

edit: divide the kbps by 1024....looks right to me

 

[NGI]

by GUI do you mean the adaptive qos page?

edit: divide the kbps by 1024....looks right to me

Yes the GUI from the Adaptive QoS page. I allways thought that the conversion from kbps to Mbps is a division by a 1000. But a Google search brings multiple results that show division by 1024 as well as 1000 so I'm confused but please ignore my question.

 

[Jack Yaz]

Yes the GUI from the Adaptive QoS page. I allways thought that the conversion from kbps to Mbps is a division by a 1000. But a Google search brings multiple results that show division by 1024 as well as 1000 so I'm confused but please ignore my question.

this is a bit of a joke but: https://xkcd.com/394/

 

[NGI]

this is a bit of a joke but: https://xkcd.com/394/

That explains it and it doesn't make me feel old school since it has been a looong time since I learn these things at uni.

 

[dave14305]

Yes the GUI from the Adaptive QoS page. I allways thought that the conversion from kbps to Mbps is a division by a 1000. But a Google search brings multiple results that show division by 1024 as well as 1000 so I'm confused but please ignore my question.

The QoS page takes Mb/s as input and stores nvram in kb/s (* 1024). But the underlying QoS tc structure takes those kb/s and applies them using 1000 (SI) instead of 1024 (IEC) bits per kilo.

 

[Jeffrey Young]

this is a bit of a joke but: https://xkcd.com/394/

Nice

 

[NGI]

Hi @Jack Yaz , I'm getting some unexpected results with ABw setting the DL speed in NVRAM so I will welcome your help if you could have a look at it.

You can see from the attached that the last scheduled speedtest at 18.42 had DL of 23.61 Mbps and 95% of it is 22.4295 Mbps but the value in the QoS page is 20.62597656 Mbps which is outside of the 5% update threshold I setup so it should have been undated. I'm not sure if it did nto get updated or it get update with a wrong value but I also can't match it with any of the previous values of the scheduled speed-tests. Am I missing something?

Many thanks in advance.

PS I have v4.1.0 and also have done force update

 

[Jack Yaz]

Hi @Jack Yaz , I'm getting some unexpected results with ABw setting the DL speed in NVRAM so I will welcome your help if you could have a look at it.

You can see from the attached that the last scheduled speedtest at 18.42 had DL of 23.61 Mbps and 95% of it is 22.4295 Mbps but the value in the QoS page is 20.62597656 Mbps which is outside of the 5% update threshold I setup so it should have been undated. I'm not sure if it did nto get updated or it get update with a wrong value but I also can't match it with any of the previous values of the scheduled speed-tests. Am I missing something?

Many thanks in advance.

PS I have v4.1.0 and also have done force update

AutoBW calculations should go to syslog

 

[NGI]

AutoBW calculations should go to syslog

Thanks Jack for clarifying that. Below is the syslog:

Nov 30 18:42:00 RT-AC86U-4AB0 spdMerlin: Starting speedtest using Structured Communications (London, United Kingdom) for WAN interface
Nov 30 18:42:20 RT-AC86U-4AB0 spdMerlin: Speedtest results - Download: 23.61 Mbps (data used: 28.0 MB) - Upload: 5.32 Mbps (data used: 2.4 MB)
Nov 30 18:42:20 RT-AC86U-4AB0 spdMerlin: Connection quality - Latency: 14.77 ms (0.19 ms jitter) - Packet Loss: 0.0%%
Nov 30 18:42:22 RT-AC86U-4AB0 spdMerlin: Calculated Download speed (22022) Kbps does not exceed 5%% threshold of existing value (21121 Kbps)

I see where is the discrepancy, I thought that the calculated speed should have been 22,968 Kbps (23.61 Mbps*95%*1024) as oppose to 22022 Kbps. Is that correct?

 

[Jack Yaz]

Thanks Jack for clarifying that. Below is the syslog:

Nov 30 18:42:00 RT-AC86U-4AB0 spdMerlin: Starting speedtest using Structured Communications (London, United Kingdom) for WAN interface
Nov 30 18:42:20 RT-AC86U-4AB0 spdMerlin: Speedtest results - Download: 23.61 Mbps (data used: 28.0 MB) - Upload: 5.32 Mbps (data used: 2.4 MB)
Nov 30 18:42:20 RT-AC86U-4AB0 spdMerlin: Connection quality - Latency: 14.77 ms (0.19 ms jitter) - Packet Loss: 0.0%%
Nov 30 18:42:22 RT-AC86U-4AB0 spdMerlin: Calculated Download speed (22022) Kbps does not exceed 5%% threshold of existing value (21121 Kbps)

I see where is the discrepancy, I thought that the calculated speed should have been 22,968 Kbps (23.61 Mbps*95%*1024) as oppose to 22022 Kbps. Is that correct?

The AutoBW feature currently uses the average of the last 10 speedtests, not the last result.