Firmware v2.5.1

CrimpOn
Guru
Dec 17, 2019
SW_ wrote:
The inclusion of "Fing" might have made the difference. In term of performance,
Fing is a popular software tool for scanning networks and deducing the manufacturer of the ethernet interfaces. It has (literally) nothing to do with speed. Speed improvements have come somewhere else.
FURRYe38
Guru
Dec 17, 2019
Been using FING for years how. Nice little app to find out whats connected on your network or find unknown devices. Hopefully won't have to use it as much since the device detection seems way better now.
SW_
Prodigy
Dec 17, 2019
CrimpOn wrote:
SW_ wrote:
The inclusion of "Fing" might have made the difference. In term of performance,
Fing is a popular software tool for scanning networks and deducing the manufacturer of the ethernet interfaces. It has (literally) nothing to do with speed. Speed improvements have come somewhere else.
Based on the incremental FW change/description, if performance benefit isn't coming from the inclusion of "Fing", i.e., enhanced devices recognition, where is this "somewhere else" exactly? For references, I don't use DIL, Armor, and Sonos. CPU uage in v2.5.0.40 is extrememly high due to devices scanning and I don't do speed test while trying to rename devices at the same time. Take a look at the picture bellow from a previous post regarding performance degradation/issue with v2.5.0.40.

Based on your above statement/claim, please share with us what you believe to be the "source" for the improved performance in v2.5.1.8. I'm ready to listen and learn about your insight. Thanks!

New Features and Enhancements:
Enhances the device recognition mechanism of the attached devices list
Add DAL into firmware for several features, like DIL, Armor.
Include Fing for device recognition.
Bug Fixes:
Fixes the compatibility issue with Sonos devices.
Update the web server certification of RBR50/RBS40
CrimpOn
Guru
Dec 17, 2019
Didn't mean to offend anyone. I have no idea what Netgear did to improve performance. What I DO know is that one person has looked at Netgear's code and found ways that he claims makes it run faster:
https://www.snbforums.com/threads/custom-firmware-build-for-orbi-rbk50-v-2-5-0-42sf-hw.60308/

Things like using the Neon instruction set, changing compiler options, etc. are open to Netgear just like they are to him. I am also happy to admit that Netgear's home grown device lookup software may have been so terrible that incorporating Fing software made a dramatic improvement in performance. Maybe Fing is "the answer" all by itself.
FURRYe38
Guru
Dec 17, 2019
Curious, are you seeing the same CPU load now with v1.8 loaded? Or is it lower?

SW_ wrote:

CrimpOn wrote:
SW_ wrote:
The inclusion of "Fing" might have made the difference. In term of performance,
Fing is a popular software tool for scanning networks and deducing the manufacturer of the ethernet interfaces. It has (literally) nothing to do with speed. Speed improvements have come somewhere else.
Based on the incremental FW change/description, if performance benefit isn't coming from the inclusion of "Fing", i.e., enhanced devices recognition, where is this "somewhere else" exactly? For references, I don't use DIL, Armor, and Sonos. CPU uage in v2.5.0.40 is extrememly high due to devices scanning and I don't do speed test while trying to rename devices at the same time. Take a look at the picture bellow from a previous post regarding performance degradation/issue with v2.5.0.40.
gr8sho
Virtuoso
Dec 17, 2019
SW_ wrote:
The inclusion of "Fing" might have made the difference.
Hello, Where did this statement come from, specifically Netgear? First I've heard of it.
FURRYe38
Guru
Dec 17, 2019
FING is mentioned in the release notes:
https://kb.netgear.com/000061530/RBR50-RBS50-Firmware-Version-2-5-1-8
SW_
Prodigy
Dec 17, 2019
CrimpOn wrote:
Didn't mean to offend anyone. I have no idea what Netgear did to improve performance. What I DO know is that one person has looked at Netgear's code and found ways that he claims makes it run faster:
https://www.snbforums.com/threads/custom-firmware-build-for-orbi-rbk50-v-2-5-0-42sf-hw.60308/

Things like using the Neon instruction set, changing compiler options, etc. are open to Netgear just like they are to him. I am also happy to admit that Netgear's home grown device lookup software may have been so terrible that incorporating Fing software made a dramatic improvement in performance. Maybe Fing is "the answer" all by itself.
It's possible that NG would suddenly decide to change the compiler options between v2.5.0.40 and v2.5.1.8 due to significant performance improvement. Assuming the compiler options were the "source", I would expect to see that reference/change description in the release notes. For example, let's take a look at the release notes for a prior release v2.3.1.44 as a reference.

-------------
v2.3.1.44

Includes bug fixes for stability and performance:
Fixes the LAN/WAN bidirectional throughput on the DHCP mode drop issue.
Fixes the downstream drop issue when port forwarding is enabled.
Fixes the issue where some online games periodicity lag.
Fixes the issue where the PPPOE connection unstable
-----------

To see the performance effect of "Fing" in v2.5.1.8, just check the speed of the "Attached Devices" page between the two releases. The devices are populated much faster and that page is more responsive with v2.5.1.8 than v2.5.0.40.
SW_
Prodigy
Dec 18, 2019
gr8sho wrote:
SW_ wrote:
The inclusion of "Fing" might have made the difference.
Hello, Where did this statement come from, specifically Netgear? First I've heard of it.
If you're referring to "Fing" within the statement, check v2.5.1.8 release note. If you're wondering about the rational behind that statement, read my posts in this thread. If in doubt, perform some tests and draw your own conclusion/statement and share the result with us.
FURRYe38
Guru
Dec 18, 2019
Voxel does his own development on some NG FW. Yes he's just started working on Orbi since he was able to get his own Orbi HW. Users have been asking for his help in the FW realm for a while now. I recommend not confusing what he does with what NG does. Stock FW differs from Voxels FW. Voxels FW is tuned and updated for outside of the NG core code. Items that Voxel has access too and is free to develop as he chooses. There are some items with in the NG code, in fact all NG code that isn't GPL, Voxel has no access or ability to change. Any problems here, NG still needs to address these here.

Read the release notes of what Voxel does. You can see that he mostly updates the FW packages that work with in the FW he has access too. Ensures that what he does works, then posts it out for people to try and use. He always open to bug reports from users to that help him fine tune any issues he may have not seen.

So yes, Voxel can help improve FW and most of the time does and has great success. His long time development of the R7800 and R9000 has helped user keep these two routers going for users who have chosen to use his FW over stock. Now that he has a Orbi 50 series system, this will benefit this series system and help users enjoy the Orbi for a more sustained longer life of operation who want to give Voxels FW a try.

Since NG has fixed the device naming issue will be nice to see Voxels version based on v2.5.1.8.
CrimpOn wrote:
Didn't mean to offend anyone. I have no idea what Netgear did to improve performance. What I DO know is that one person has looked at Netgear's code and found ways that he claims makes it run faster:
https://www.snbforums.com/threads/custom-firmware-build-for-orbi-rbk50-v-2-5-0-42sf-hw.60308/

Things like using the Neon instruction set, changing compiler options, etc. are open to Netgear just like they are to him. I am also happy to admit that Netgear's home grown device lookup software may have been so terrible that incorporating Fing software made a dramatic improvement in performance. Maybe Fing is "the answer" all by itself.
gr8sho
Virtuoso
Dec 18, 2019
Thank you.

When looking at the information presented for scope of the update, all that was readily available was stated in the OP. Not sure why the linked release notes weren’t more prominent . Strange manner of communication.

As for performance claims, that’s fine as it hasn’t been an obvious issue here.

The only other thing I’ll add is the IGMP support continues to be a fail unfortunately.
Hanseman
Tutor
Dec 18, 2019
Has anybody noticed better values for dBm since this new firmware?
Perhaps a coincidence but I measure this regularly and the connection between router and satellites are more stable and dBm values has improved to both satellites (and they have not been moved).
I was planning on moving the router closer to the satellites but now I'm waiting a while.
DaVincee
Tutor
Jan 09, 2020
I posted my two cents about device renaming before finding this ariticle and the new firmware. Applying it today and see what happens. Thanks for the heads up on the forums: it makes life a bit more predictable :smileyvery-happy:

Forum Discussion