[yakumo]

386.02_4-gnuton1  – cake does not work, easily proven by dslreports buffer bloat test showing results into the hundreds of ms. Gnuton does know, it’s raised on github here. Lots of nice additions though, and holding over 200mb free ram so far.

 

While testing this I also found that if you had Traditional QoS rules set up  and then either turn QoS off or swap to another type then when you return to Traditional even though your previous rules may show you’ll find tc qdisc via ssh will not show fc_codel until you go to the rule page and hit apply again. I would not be surprised if this was the same with the stock firmware, but I’m not flashing back just to check.

Share the knowledge

You need to login in order to vote

[yakumo]

That only leads to people having to install a program to generate checksums themselves to check the one given, and Sentinel being asked about that by confused non-tech users.
It’s infinitely simpler to just to zip the file, it’s opened in windows when you click it without any need for additional software, if it doesn’t let you drag the file out there is a problem.

Please forget that I suggested 7z for here at all, it’s just a better format that I’m fond of encouraging use for, also due to it’s comprehensive format supporting free ui, counter to seeing even companies using winrar decades beyond it’s trial. But it’s just overcomplicated the conversation.

Share the knowledge

You need to login in order to vote

[yakumo]

Grisu wrote:

You can make a CRC check with any file, 7-zip give you the option to run multiple CRC checks. Only thing you should provide checksum.

It’s inherent to the 7z and zip format, all the files in the archive always get CRC32 checksums, all files are checked when uncompressed and an alert shown if any fail.

Share the knowledge

You need to login in order to vote

[yakumo]

That’s a valid reason to not use 7z despite it’s many year gaining popularity and completely free client 7-zip, but standard zip support has been completely integrated in the Windows File Explorer since 1998.
And though I don’t wish to come across as tech-elitist – anyone that does not know what a zip file is should really not be installing firmware from anywhere but the manufacturer, though many of those supply all of their files zipped as it is (including Asus).

Share the knowledge

You need to login in order to vote

[yakumo]

3.0.0.4.386_42438
The QoS – WAN/LAN Bandwidth Monitor when App analysis is off stops working sometimes, refreshing repeatedly, logging in/out doesn’t solve this, I tried a few times over an hour, but it seems to have righted itself without a reboot by the next day on the last occasion. I’ve noticed this often happens if app analysis has been turned on/off quite a few times (short progress bar caused), but the last time it occurred was after altering a QoS priority rule and applying.

DSL-AX82U_9.0.0.4_386_42943-g4e3dfd3_cferom_puresqubi.w
– could firmware uploads here be 7zipped or at least zipped  (7z ~4mb better filesize on the last beta), there is no checksum with it so it’s not possible to be sure the file is intact prior to attempting install otherwise, both zip formats have error checking. Cheers.
– searching for the “DSL-AX82U_9.0.0.4_386_42943” I came across a Turkish forum on donanimhaber, thanks to google translate I see a discussion of poor LAN throughput from one user and poor DSL speeds from another  with that build and 9.0.0.4_376_42258 as well, they rolled back to 3.0.0.4_386_40137, your mileage may vary but I thought I’d give a heads up.

Share the knowledge

You need to login in order to vote

[yakumo]

Well, this is why I would not have tried this if there were not going to be updates or custom firmware. I would like to try cake

Share the knowledge

You need to login in order to vote

[yakumo]

If you are discussing things with Asus, their QoS FAQ point 9 is not very clearly written though top down priority overruling any lower rule is still conveyed, but far more importantly raises the question of – with order so important why is it not yet possible to rearrange the order?

In the worst case scenario wanting to alter any of the fields for the upper most rules other than the priority drop down results in having to rebuild the entire stack.

Share the knowledge

You need to login in order to vote

[yakumo]

> manually allocating a set bandwidth per device is good but is time consuming .

I was testing altering bandwidth percentages by QoS priority rather than per device, but I think a slightly higher reduction on the overall bandwidth setting is a better solution, leaving the priorities at default. A FireTV Echo is causing real problems at times when in use with zero to 36Mbps in regular 3 second pulses even if set to a 20mbps device limit, it is more WAN disruptive than any P2P so far, if I can’t find a solution by router or by it’s own settings it may need replacing with a Shield.

Diabotical (free) and Quake Live are UDP based and excellent tools for testing latency. They have network graphs showing real time packet flow and packet loss and they’re very sensitive, sampling at a higher rate than ping/trace tool defaults, higher than many allow. As the latency is measured via the game’s usual UDP datagram, there is no risk of it being artificially shaped anywhere on route separately to the game traffic.

With Diabotical certain network conditions can also cause the game to apply automatic network restrictions, helping to prevent warp or abuse spoiling gameplay for other players, when these restrictions are in play the graph background turns red. The exact conditions that trigger this have not been specified but this is a third quality metric which can show even if latency and packet loss otherwise appear to be unaffected.

Testing involved playing while other consistent sources of traffic were seen or intentionally generated on the network, including SSL, OpenVPN, Netflix and Amazon streaming, a bittorrent of the latest iso from https://ubuntu.com/download/alternative-downloads .

Testing on an 18ms game server, the Adaptive QoS gave consistently higher latency in all heavy load test cases typically by 50ms or more and lost packets, the automatic network restriction was almost constant, with Traditional it was possible to eliminate that almost completely rarely gaining more than 7ms under heavy load (Ubunutu P2P) and without triggering the network auto restrictions, even if it they flared up for a moment on a large network spike it cleared very rapidly.

I need to investigate bufferbloat testing more, but dslreports speed test has a bufferbloat stat that was +200ms or more when using Adaptive QoS, if the experience in game had been better while the test was running I would have assumed it was entirely due to game traffic taking too much of a priority over the web, but the game was unplayable and the bufferbloat report also remained as poor when no game traffic was present so the QoS rule should have remained untriggered.

With Traditional rules and simply setting known game UDP ports to highest priority and the dl and ul main configuration set to 92% for sufficient reserve there is no need to alter the default bandwidth priorities and game experience is nearly flawless except under the most extreme loads, Adapative is simply not suitable for realtime gaming at all.

The Bufferbloat stat with this set up remained 2-5ms almost entirely after an initial spike at the start of the DSLreports traffic calmed.

TCP based games would not highlight problems as well due to TCP’s inherent ordering, retransmission, error recovery, only latency would remain an issue, but it is not often used for fast paced realtime games.

Thank you for the note on  fq_codel, I was wondering if something like that may be the case, now that is known I don’t need to experiment with custom firmware just yet.

Share the knowledge

You need to login in order to vote

[yakumo]

I’ve had the ax82u since Sunday, so far very happy, always at least 140mb ram free whatever I have experimented with (currently traditional QoS & app analysis on over the last day to help tune QoS ) so not concerned about memory so far. It does seem to be handling house traffic better than the 88u, but that only allowed for upload limiting and gave so little information that could be used to balance properly.

Plugging it in called for an update to 3.0.0.4.386_42438 as soon as it was online. It’s connecting to the ISP a lot faster than the ac88u which is interesting, but no real change on line stats so far. I’m guilty of hoping it might increase max rate eventually, but the line itself is old and remote so it may be the best it will ever maintain stability on.

For anyone with low latency/low bufferbloat needs : Traditional QoS with few or many rules and manually set bandwidth for the initial QoS setup still seems to win out over Adaptive whether bandwidth is auto or manually set, maintaining lower latency overall when testing UDP games.
Without deeper investigation The Game Gear Accelerator seems to only be a simple UI for adding your systems into Adaptive QoS as highest priority, Adaptive is switched to if you enable Gear Accelerator, it certainly does not appear significantly better than Adaptive enabled via other means, but I did not look into checking the exact config via telnet to see if there were any real configuration differences, maybe another time.
Traditional rules are not deleted so you can swap easily between the two if you wish to test.

Enabling Adaptive (or Gear) from Traditional causes a quick progress bar (<10s) which did not noticeably terminate traffic, going back the other way appears to cause a reboot however, or at least cut traffic for well over a minute. Considering the most valuable use of QoS is managing a multi-user network, I really don’t think it’s acceptable such operations aren’t always proceeded with a clear warning and opportunity to back out, especially as this example shows it’s not entirely predicable, that goes for any operation that will cause a disconnect really, not just QoS changes.

I’m going to pass on installing beta firmware for now unless there is a very compelling reason, and hope some of the new updates hit stable status soon.

Share the knowledge

You need to login in order to vote

[yakumo]

The dsl-ax82u price from Amazon co.uk (direct) just dropped £45, it’s not a price guarantee pre-order item so you’re not updated automatically while they wait for stock, I was lucky to spot it, and was able cancel and re-order, ETA window unchanged.
So if anyone else has ordered you might want to check.

Share the knowledge

You need to login in order to vote

[yakumo]

UK Sentinel wrote:

I remember from the DSL-AC68u days, I encouraged a neighbour  to purchase the DSL-AC68U and after some disconnects and calling out out a BT OPENreach Engineer, my neighbour via the engineer was asked to remove the DSL-AC68U ?

Yes, that specific unit ended up getting blacklisted by bt/openreach at one point for an option it had available that was causing network problems for others on the same cabinet, the Mediatek modem chip it used had severe issues after the g.inp upgrades had rolled out and unless it caused the occasional disconnect the owner usually had no knowledge there was a problem but it caused interference for other customers on the same cabinet. Once flagged by the system it only became impossible for it to connect with g.inp and it’s performance was heavily capped by DLM rather than being blocked from connection to the network at all,  I had one at the time and went through this myself, I expected to be told to remove it if I hadn’t swapped it out as soon as I saw the discussion of it being blacklisted online, however fighting to try and get it back up to speed with different configurations I did specifically mention it was in use in multiple calls to BT as fastpath dropped out a few times over months but I was never told to actually stop using it, as soon as I was able though I swapped to a Huawei modem instead and had full speed and g.inp within 30 minutes without having to wait days for DLM or again request a reset.

The engineer you heard of may have just suggested removing what was an unknown router to them as the customer was complaining of disconnects, use the stock router is default advice, or there is a chance they recognised it specifically by then and knew it was best removed. If the situation had actually escalated further in the 2 years I kept the device as a bridge I did not see it mentioned on the forums though.

With the SIN review above giving me confidence about it’s actual standing, I am going to give the 82u a try as the LAN & W-LAN side of things should be excellent still, and I like the firmware a lot over other brands. The firmware licensing not being an issue any more is a huge plus that I hope will result in updates & custom patches for years.

I’m definitely not going through the MediaTek fiasco again or suffer firmware stagnation, if there are any issues not resolved very quickly then I will swap it out for a fritzbox immediately on your recommendation and just have to cope with any loss of hw/sw features, unless by then there is something clearly better from any upcoming refresh due to 6E wifi but they could easily price themselves out of the equation, I would wait for those now but for wanting rid of the ac88u’s closed fw and 6E unit’s expected price hike.

Share the knowledge

Liked by:

You need to login in order to vote

[yakumo]

Yes, I saw that, but nothing in the actual SIN 498 MCT document provided by Openreach appears to actually back up that particular quote including the definition of a “SIN” document itself (Suppliers’ Information Note) unless you are a  Communications Provider (AKA. Supplier), which makes it seem like there has been some miscommunication within some manufacturers, or they’re running ahead on information which does not actually seem to have been documented yet while pointing to what would then be out of date documentation as the source (a very strange source considering it’s title classification).

Without being able to find a statement or document from BT/Openreach themselves declaring this status change overruling the content of SIN498 itself, it honestly looks like an internal company memo has gone around (in 2016) warning of document compliance becoming a requirement while accidentally omitting exactly who that requirement pertains to, and then it’s fallen into a helpful advertising bullet point which is still fairly factual, but not quite entirely genuine for end users if the SIN 948 MCT is taken as-written by itself.

*other than the simple generic single instance of  ‘if your equipment disrupts our service then it will be blocked’ which would normally be assumed as obvious anyway.

“I have decided you must work for ASUS”  no but if they would happen to offer consultant fees I’d be happy for them to get in touch, hard times after all ;-)

Share the knowledge

You need to login in order to vote

[yakumo]

I have gone through the SIN 498 MCT and I believe there have been some misunderstandings spreading online for some time about it –

Obligatory; I Am Not A Lawyer But –

<!–more–>

“This Suppliers’ Information Note (SIN) provides details relevant to Communications
Providers (CPs) regarding connectivity and interfaces.”

“The detailed technical requirements for CP provided modems are defined in Section 3
of this document and the related test descriptions required to demonstrate compliance
to these requirements are defined in Annex A.
The terms and conditions associated with the communication and deployment of those
network changes can be found within the GEA Contracts which are available from
http://www.openreach.co.uk/orpg/home/products/superfastfibreaccess/contracts/sffacontracts.do
CPE connected to Openreach’s network will be expected to be upgraded to remain
compliant with the evolving Openreach network, as reflected in changes to this SIN.”

CP = Communications Providers (these are not End Users, but ISP’s etc providing services on mass to end users).
CPE = Customer Premises Equipment (this is in the context of having been supplied by the CP in all except one instance noted later)
EU = End Users  (The customer purchasing the connection for their abode)
SIN = Suppliers’ Information Note
MCT =  Modem Conformance Test

“1.2.1
…
CPs should also be aware of the following:
 Where Openreach has provided the modem, daily status reports will be
generated and transmitted consisting of no more than 8k bytes (64k bits) of
data upstream at full line rate. These flows will take priority over EU data. The
impact clearly depends on the VDSL2 traffic rate at the time”

 

“2.1.3.2 CP added tags

…
EU CPE such as set top boxes (STB) and PCs may add X-tags in the upstream
direction and these will be transported transparently through to the CP. An
exception to this is tag 0 which will be removed by Openreach (see section 2.2.3
– Upstream priority marking – for more detail).”

The quote immediately above is relevant as it is the one and only time “EU CPE” appears in the entire document. There is also no occurrence of any longer form such as “End User CPE” or the initialism being written out completely in full.

“2.1.8 Intermediate Agent / DHCP Relay Agent
Where PPPoE is detected, additional tags will be inserted into the upstream flow
(PADI) by the Intermediate Agent (IA) in the DSLAM. Any existing tags of the same
type from the CPE will be overwritten. The IA tags will be removed by the DSLAM
in the downstream direction (i.e. from the PADO, PADS messages).”

 

“2.2.4 Modem UNI Port Loopback Testing
Test and diagnostic action may require an Ethernet port loopback to be applied to the
modem VDSL2 port in order to loop downstream traffic back upstream to the
Openreach test head or CP test head. These tests will interrupt upstream traffic from
the EU and should therefore only be enabled with the EU’s consent. The EU must also
agree to stop any downstream Multicast Service traffic and power off their Set Top Box
if they have one, as any”

 

“2.4 CP Provided Modem Product Variant
Openreach have introduced a GEA-FTTC product variant that allows the CP to provide
and be responsible for the user’s VDSL2 modem. Typically, this modem will be
integrated with IP gateway functionality within a single device and connected to a
single mains power source. CPs or their EUs will be responsible for maintaining the
firmware of their modems and monitoring their connectivity and performance, typically
via a TR-069 [19]
interface using CPE WAN Management Protocol (CWMP).
The CP provided modem and filtering devices must meet the requirements of this
specification in order to provide reliable operation and to avoid harm to other VDSL2
lines sharing the same cable binder. Openreach reserves the right to withhold or limit
service where potential violation of the Access Network Frequency Plan (ANFP) [9]
or impact to another customers’ service is detected.“

They reserve the right to remove your service if your equipment is causing problems with their network or other customers, this would apply to any equipment and is a generic catch all for any kind of service that is highly common in T&C’s or EULAS, allowing them to ban you at-will for behaviour they deem harmful or disruptive. However, this does not require the EU’s own equipment to have passed the certification defined by this document in the first place, that requirement only appears to be if it is CP provided equipment (for mass distribution).
As such this is not really any more demanding for an End User than if their completely SIN 498 MCT certified modem developed a fault, had a bad flash or they reconfigured the hardware or software themselves out of spec.

Hopefully this will put some at ease a little as so long as there is no actual disruption from your modem (stability adjustment settings could cause this, defaults should be fine), there is no grounds for it to be removed simply for being discovered on the line as a non-certified product, you are not bound by the rest of this Communications Provider oriented agreement and required to be using a certified product as they are.

As such I can completely understand why Asus has not been applying for SIN 498 validation, they’re not seeking distribution via CP’s, and are ok with handling any issues that come up themselves, or their users facing what would almost certainly only be temporary restrictions or removal from the network until the issue was resolved (though that could in the worst case mean replacing the device if it couldn’t be made to comply).

<!–more–>

“The Communications Provider shall mark the Openreach termination point on a fire
retardant wooden backboard with a footprint of 90mm x 115mm with allowance of at
least 45 mm for cable entry into the bottom edge of the termination”

This just further reinforces that Communications Providers are not the End User’s / aka general public buying the service.

“2.9.1
…
For compatibility reasons, CPs providing their own modem CPE are also
required to implement this functionality in their modem devices.”

“3. CPE Requirements For GEA over VDSL2
This section defines the requirements of CP provided modems that must be met for
connection to Openreach User Network Interface (UNI). These requirements include
logical functions within the CPE necessary to support and maintain Openreach services
delivered over GEA-FTTC”

 

“4. CPE Requirements For FTTC GEA over ADSL2plus (OPTIONAL)
This section defines the requirements of CP provided modems that must be met for
connection to Openreach User Network Interface (UNI) for FTTC GEA over
ADSL2plus. These requirements include logical functions within the CPE necessary
to support and maintain Openreach services delivered using FTTC GEA over
ADSL2plus.”

 

“Annex A Test Requirements for GEA over VDSL2
This Annex provides a detailed breakdown of the modem conformance test (MCT)
requirements to enable a piece of vendor CPE to be validated against the CPE
Requirements defined in Section 3 of this document.”

Please take care not to quote this entire gigantic post if replying ;-) I am sorry the (more) tags offered by the toolbar do not appear to work.

Share the knowledge

Liked by:

You need to login in order to vote

[yakumo]

Thank you, I’ve seen you post over many years now and value your experience.

 

Share the knowledge

Liked by:

You need to login in order to vote

[yakumo]

Strange, I had wanted to find current details but 498 is missing from https://www.bt.com/about/sinet/sins/downloads
Edit: found it on Openreach

Any particular stand out products for an alternative with fairly equivalent hardware and QoS features?

Share the knowledge

You need to login in order to vote

[Author]

Posts