What could cause these packet loss spikes?

Hey all

I will go through what has happened and what I have done first and then at the end ask my question(s).

I have moved away from the Smart Hub due to my one doesn't work with IPv6, so I brought a FRITZ!Box 4040 Router.

This is to be connected to my ONT via WAN and has been working fine for a few days.

Then a couple of days ago I started getting about 10% packet loss around every 6 hours.
So it spikes just before 4am, just before 10am, just after 4pm and just after 10pm, I have absolutely no clue what is causing it.

Here is an image where you can see it spiking, the latency spike on the right was me doing a speed test.

The only thing that has changed from when it was all fine was the Dynamic IP, maybe that IP had some issues.

After speaking to BT Care via Twitter I a full reboot of the router and disabled the old BQM and setup a new one with the new IP and left it going which is shown here.

The latency spike on the left was me doing a speed test and the packet loss spikes was me changing some setting which for some reason dropped the connection.

Now the stupid part:

Sadly due to viewing the wrong BQM Graph I thought it was still doing it, so I did another reboot, sadly I got a new IP, so had to go through the same process with the BQM
So now the third BQM in my Sig, that little red spike was me locking down the router to allow just ICMP (Stealth Mode)

*** Question(s) ***

So my question, what would cause those packet loss spikes at the same times?

I am on an Infinity 4 connection where absolutely everything is over fibre, so am I correct it was an issue BT's end to do with that IP I was given for those two days that I was having the issue?

I ask this due to all the other IPs that I have been given on this router have not had any packet loss.

I have made a note of that offending IP so if I get it again in the future I will just reboot the router and get another IP.

Paul

First image with the big lumps of red looked very regular almost as if a router was delaying responses for a while.

In terms of the yellow generally looks very healthy, assuming you are sometimes using the connection

We need to see a few hours of BQM3 of course .

Pure speculation -
How long had the router been up when the issue started?
Maybe after some time it will happen again and a reboot will clear it.
What is the router?
Edit: Doh ignore the last question!

Edited by deleted (Fri 21-Apr-17 10:59:33)

In reply to a post by MrSaffron:

First image with the big lumps of red looked very regular almost as if a router was delaying responses for a while.

In terms of the yellow generally looks very healthy, assuming you are sometimes using the connection

Yeah, that's why its confusing me, BT Care said I would need to connect my Smart Hub when/if I phone the FTTP Support Team, but with the Smart Hub I would be unaware if the issue is still there or not due to it blocks ICMP Packets.

Like I said it was fine for the first few days when on a different IP.

I did some changes to-do with Wireless where I then rebooted the router this was where I was given a new IP which had the 10% packet loss around every 6 hours.

This went on for a further 2 or so days like this until I rebooted yesterday where I did another reboot, got a third IP Address and was fine again.

I did no changes this time, so what caused the 10% packet loss before should still be there, but it wasn't, so I was assuming it was linked to that IP.

I did tell BT Care via Twitter that I will leave it for another 24 to 48 hours to see if it happens again, but the one that run for the last 21 hours was fine.

Paul

In reply to a post by RobertoS:

We need to see a few hours of BQM3 of course .

Yeah, I will be leaving it for the next 24 to 48 hours, I am sure it will be fine and that bad couple of days was linked to the IP I was given that time.

Paul

In reply to a post by legume:

Pure speculation -
How long had the router been up when the issue started?

Well it was setup on the 6th where it was fine, I just had issues with the wireless, so I did some changes to that and I was told I would have to do a full reboot for those changes to take affect, which I did.

Those few days from the 6th was fine until the reboot on the 10th, I got a new IP and everything was very slugish when browsing, I would also get DC'd from games at random times.

I did a reboot yesterday and everything was fine again.

In reply to a post by legume:

Maybe after some time it will happen again and a reboot will clear it.
What is the router?
Edit: Doh ignore the last question!

Well we won't know until it happens, my guesses are that the issue was down to a bad IP, maybe a routing issue due to an intermittent fault.

Its just strange that nothing was changed to-do with the firewall, or the internet settings, as far as I know as soon as it was given that bad IP I got the issue and after a reboot I was given a new IP and it was fine.

But I will leave it running for the next 24 to 48 hours and see if it comes back.

Paul

The IP change indicates probably that your routing through the BT Consumer network may have changed, i.e. nothing to do with router or your FTTP.

FTTP only solves the attenuation and noise issues of copper, it does not guarantee a perfect core network

In reply to a post by MrSaffron:

The IP change indicates probably that your routing through the BT Consumer network may have changed, i.e. nothing to do with router or your FTTP.

FTTP only solves the attenuation and noise issues of copper, it does not guarantee a perfect core network

That's what I am trying to say, all other IP's that I had apart from that one which I had the issues with were fine, so that's why I said I think I must of had an IP that had bad routing or some issues with it.

So far the current BQM has gone passed the 10am mark with no issues, so maybe this is ok now.

I will actually leave it for a week and check it once a days to see, but I am sure it will be fine.

Paul

In reply to a post by MrSaffron:

The IP change indicates probably that your routing through the BT Consumer network may have changed, i.e. nothing to do with router or your FTTP. FTTP only solves the attenuation and noise issues of copper, it does not guarantee a perfect core network

Reminds me of the days of dialup when you'd disconnect and re-dial to get a different connection; normally the same on the modem side, and different routing from a different RAS bank at the ISP.

Well its seems to of not happened yet, but I will leave it for another 6 days.

Paul

Well its back again and it was worse this time shown here, not really sure what to do now.

I will leave it going for another few days and then decide.

Anyone know of a good router to be used on Infinity 4 via the WAN Port, it needs to support IPv6 and respond to ICMP Pings?

Paul

Edited by PaulKirby (Thu 27-Apr-17 23:22:21)

An ASUS perhaps? Their router and wifi side seems good, built-in modems that you aren't interest in a different matter.

In reply to a post by RobertoS:

An ASUS perhaps? Their router and wifi side seems good, built-in modems that you aren't interest in a different matter.

Maybe, I am currently looking at a Linksys WRT3200ACM-UK on Amazon, I did look at others but I either didn't like the look of it or it had bad reviews.

*** update ***

Well I just ordered a Linksys WRT3200ACM-UK Router and should get it Saturday, but I will leave the Fritz 4040 connected for the next couple of days to see what happens, I think it will keep spiking around every 6 hours.

So this new router should rule out either the other router or the connection, just hoping its not the ONT LOL.

Paul

Edited by PaulKirby (Fri 28-Apr-17 03:24:20)

Nahhh! It's not the ONT. It's the position of the moon over 24 hours in relation to your house and the route to the ISP routers. Like the tide at every particular beach. It's pulling on a tiny manufacturing fault in the fibre every six hours and causing the problem.

You can see this from the shape of each episode.

In reply to a post by RobertoS:

Nahhh! It's not the ONT. It's the position of the moon over 24 hours in relation to your house and the route to the ISP routers. Like the tide at every particular beach. It's pulling on a tiny manufacturing fault in the fibre every six hours and causing the problem.

You can see this from the shape of each episode.

LOL.

It is strange though isn't it.

If it was noise it would be the same no matter if I rebooted / reset the connection, however its fine for X days then at an interval of the time it reconnected and every 6 hours after.

This time it was rebooted / reset on 21.04.2017, 07:44 o'clock and one of the spikes is around that time.

Where as before when it was happening the connection was rebooted / reset on 10.04.2017, 03:49 o'clock and it was happening around that time and every 6 hours after that.

TBH I don't think its the ONT due to that's not been touched since it was installed.

Its just the X days after the connection is reset where it is fine before the spikes happen is confusing resulting in not knowing if its the router or the connection.

I am currently on the fence here due to this FRITZ|Box 4040 is rather a new product and isn't really sold to the UK hence there not being a UK plug on the PSU, also I am having issues with the Wi-Fi where its dropping connections so that doesn't help much.

So hopefully this new Linksys router will either show its the FritzBox or the connection, plus the Wi-Fi cannot be as bad as this current one LOL.

Paul

Noise as in photons being corrupted?

Remember FTTP is not subject to RF noise.

The fact its showing packet loss with no obvious change in baseline latency tends to point to a broadband hardware issue

In reply to a post by MrSaffron:

Noise as in photons being corrupted?

Remember FTTP is not subject to RF noise.

Yeah, I am aware of that, but BT still say that.

In reply to a post by MrSaffron:

The fact its showing packet loss with no obvious change in baseline latency tends to point to a broadband hardware issue

What hardware would this be, BT's internal hardware, the ONT, or my Router.

And like you said the latency is still very low when its happening, which is confusing me.

Also it doesn't happen right away, it starts happening X days after the new connection is made, and the one of the 6 hour spike will be the time it was reconnected on.

BT has told me via Twitter that I would have to connect the Smart Hub back up so that they can do they tests, that wouldn't help if it doesn't happen right away, also how would I know its happening due to the hub blocks ICMP Pings and there is nothing stopping BT from saying everything is fine when it might not be.

Like I have already said I have had to resort to buying another completely different branded router only product to rule out it being a router issue, its a shame that the Smart Hub doesn't have an expert mode which allows the user to allow for ICMP pings.

Granted we have the DMZ, but I don't really want to setup a server just to allow for ICMP ping responses.

Paul

Edited by PaulKirby (Fri 28-Apr-17 12:34:53)

My money is on your router

There is the question too that are you seeing anything actually impacting on the service, ICMP is designed to be a low priority protocol so when software in routers gets busy it is put on hold, and it may be that in your case the router is just dropping them periodically.

So when you see your next block you need to run some outbound ping and traceroute testing to see if this is a service impact event or just phantom

In reply to a post by MrSaffron:

My money is on your router

There is the question too that are you seeing anything actually impacting on the service, ICMP is designed to be a low priority protocol so when software in routers gets busy it is put on hold, and it may be that in your case the router is just dropping them periodically.

So when you see your next block you need to run some outbound ping and traceroute testing to see if this is a service impact event or just phantom

I was at one point wondering if it was the router due to it is very new and only become available on 26 Aug. 2016 it also has some other small issues, so another braded router might be good.

I have always missed when the actual spikes happen, but its due to happen very soon, so I might run some pings and tracroutes in a loop.

Paul

Well the spike has been and gone.

All pings to my server was 17ms at all times which is normal.

However I trace routed to here, hope you didn't mind and I have some of the results below:

tracert -4 thinkbroadband.com

Tracing route to thinkbroadband.com [80.249.99.130]
over a maximum of 30 hops:

  1    <1 ms    <1 ms    <1 ms  YAZOOROUTER [192.168.2.253]
  2     3 ms     2 ms     2 ms  217.32.146.68
  3     5 ms     3 ms     3 ms  217.32.146.94
  4     3 ms     3 ms     3 ms  213.120.156.210
  5     3 ms     3 ms     3 ms  217.41.168.49
  6     4 ms     3 ms     3 ms  217.41.168.107
  7     4 ms     3 ms     3 ms  109.159.249.104
  8     3 ms     3 ms     3 ms  core3-te0-10-0-19.faraday.ukcore.bt.net [109.159.249.23]
  9     4 ms     4 ms     4 ms  213.121.193.191
 10     4 ms     3 ms     3 ms  linx-gw1.thn.ncuk.net [195.66.224.240]
 11     3 ms     4 ms     3 ms  te2-1-9.star10g.bdr-rt3.thdo.ncuk.net [80.249.97.17]
 12     *        *        *     Request timed out.
 13     4 ms     3 ms     4 ms  www.thinkbroadband.com [80.249.99.130]

Trace complete.

tracert -4 thinkbroadband.com

Tracing route to thinkbroadband.com [80.249.99.130]
over a maximum of 30 hops:

  1     1 ms    <1 ms    <1 ms  YAZOOROUTER [192.168.2.253]
  2     3 ms     2 ms     2 ms  217.32.146.68
  3    22 ms     5 ms     8 ms  217.32.146.94
  4     3 ms     3 ms     3 ms  213.120.156.210
  5     3 ms     3 ms     3 ms  217.41.168.49
  6     4 ms     3 ms     3 ms  217.41.168.107
  7     3 ms     3 ms     3 ms  109.159.249.104
  8     4 ms     3 ms     3 ms  core3-te0-10-0-19.faraday.ukcore.bt.net [109.159.249.23]
  9     4 ms     4 ms     4 ms  213.121.193.191
 10   166 ms   213 ms     *     linx-gw1.thn.ncuk.net [195.66.224.240]
 11     *        *        *     Request timed out.
 12     4 ms     3 ms     4 ms  www.thinkbroadband.com [80.249.99.130]

Trace complete.

tracert -4 thinkbroadband.com

Tracing route to thinkbroadband.com [80.249.99.130]
over a maximum of 30 hops:

  1    <1 ms    <1 ms    <1 ms  YAZOOROUTER [192.168.2.253]
  2     3 ms     2 ms     2 ms  217.32.146.68
  3     3 ms     3 ms     2 ms  217.32.146.94
  4     4 ms     3 ms     3 ms  213.120.156.210
  5     3 ms     3 ms     3 ms  217.41.168.49
  6     3 ms     3 ms     3 ms  217.41.168.107
  7     4 ms     3 ms     3 ms  109.159.249.104
  8     4 ms     3 ms     3 ms  core3-te0-10-0-19.faraday.ukcore.bt.net [109.159.249.23]
  9     4 ms     4 ms     4 ms  213.121.193.191
 10     4 ms     4 ms     4 ms  linx-gw1.thn.ncuk.net [195.66.224.240]
 11     4 ms     4 ms     4 ms  te2-1-9.star10g.bdr-rt3.thdo.ncuk.net [80.249.97.17]
 12     *        *        4 ms  www.thinkbroadband.com [80.249.99.130]

Trace complete.

tracert -4 thinkbroadband.com

Tracing route to thinkbroadband.com [80.249.99.130]
over a maximum of 30 hops:

  1    <1 ms    <1 ms    <1 ms  YAZOOROUTER [192.168.2.253]
  2     3 ms     2 ms     2 ms  217.32.146.68
  3     3 ms     2 ms     2 ms  217.32.146.94
  4     4 ms     3 ms     3 ms  213.120.156.210
  5     3 ms     3 ms     3 ms  217.41.168.49
  6     3 ms     3 ms     3 ms  217.41.168.107
  7     3 ms     3 ms     3 ms  109.159.249.104
  8     3 ms     3 ms     3 ms  core3-te0-10-0-19.faraday.ukcore.bt.net [109.159.249.23]
  9     4 ms     4 ms     3 ms  213.121.193.191
 10   124 ms   216 ms     4 ms  linx-gw1.thn.ncuk.net [195.66.224.240]
 11     *        *        4 ms  te2-1-9.star10g.bdr-rt3.thdo.ncuk.net [80.249.97.17]
 12     4 ms     3 ms     4 ms  www.thinkbroadband.com [80.249.99.130]

Trace complete.

I am not really a network routing guy, so the above odd * is fine but the "Request timed out" isn't.

I am still no closer to knowing if its a router issue or connection.

I just re did some trace routing to here where there is no spikes and its all fine again:

tracert -4 thinkbroadband.com

Tracing route to thinkbroadband.com [80.249.99.130]
over a maximum of 30 hops:

  1    <1 ms    <1 ms    <1 ms  YAZOOROUTER [192.168.2.253]
  2     3 ms     2 ms     2 ms  217.32.146.68
  3     3 ms     2 ms     2 ms  217.32.146.94
  4     4 ms     3 ms     3 ms  213.120.156.210
  5     3 ms     3 ms     3 ms  217.41.168.49
  6     3 ms     3 ms     3 ms  217.41.168.107
  7     3 ms     3 ms     3 ms  109.159.249.104
  8     3 ms     3 ms     3 ms  core3-te0-10-0-19.faraday.ukcore.bt.net [109.159.249.23]
  9     3 ms     4 ms     4 ms  213.121.193.191
 10     4 ms     4 ms     4 ms  linx-gw1.thn.ncuk.net [195.66.224.240]
 11     4 ms     4 ms     4 ms  te2-1-9.star10g.bdr-rt3.thdo.ncuk.net [80.249.97.17]
 12     4 ms     3 ms     3 ms  www.thinkbroadband.com [80.249.99.130]

Trace complete.

tracert -4 thinkbroadband.com

Tracing route to thinkbroadband.com [80.249.99.130]
over a maximum of 30 hops:

  1    <1 ms    <1 ms    <1 ms  YAZOOROUTER [192.168.2.253]
  2     3 ms     2 ms     2 ms  217.32.146.68
  3     3 ms     2 ms     2 ms  217.32.146.94
  4     4 ms     3 ms     3 ms  213.120.156.210
  5     3 ms     3 ms     3 ms  217.41.168.49
  6     3 ms     4 ms     3 ms  217.41.168.107
  7     3 ms     3 ms     3 ms  109.159.249.104
  8     4 ms     3 ms     3 ms  core3-te0-10-0-19.faraday.ukcore.bt.net [109.159.249.23]
  9     4 ms     4 ms     4 ms  213.121.193.191
 10     4 ms     4 ms     4 ms  linx-gw1.thn.ncuk.net [195.66.224.240]
 11     4 ms     3 ms     4 ms  te2-1-9.star10g.bdr-rt3.thdo.ncuk.net [80.249.97.17]
 12     3 ms     4 ms     4 ms  www.thinkbroadband.com [80.249.99.130]

Trace complete.

Very strange.

Paul

Not strange at all, devices prioritise processing through traffic rather than processing ping/icmp requests, only if subsequent hops are higher would you need to worry.

We also monitor interfaces and if latency exceeds normal limits for a period of time will get alarms notifying us and there was none, and if it was those nodes everyone coming across that link would also be affected.

Ah, ok, so any clue what could be causing those spikes ?

I am hoping its the router due to its easily replaceable, getting BT to do something is harder to do LOL.

Paul

I wonder if it is calling home and waiting for a response with a CPU loop running? (Serious, not a joke this time). Don't forget, the dropped packets are not a measurement by your router, they are BQM not getting replies from it.

The router busy doing something else is a possibility, in software terms garbage collection i.e. recycling memory is a possible reason for the delays.

The traceroutes seem to show outbound packets and responses are not affected, supporting hypothesis that this may be a firmware issue in the broadband router

Ok, well the new Linksys WRT3200ACM-UK is down to arrive sometime tomorrow.

Then it will be a waiting game for another several days to see if it happens on this one, which I am hoping it doesn't.

Paul

Well the Linksys WRT3200ACM arrived, all set up.

Wi-Fi is great.

Sadly my 300Mbit connection is now a 4Mbit connection if I am lucky

Linksys Speed Test

Its setup as PPPoE, also set the usernsame to [email protected] same as what it was on my Fritz!Box 4040

Oh and IPv6 isn't working.

Right I am swapping back to the Fritzbox to see if that connection is better.

Paul

http://community.linksys.com/t5/Wireless-Routers/WRT...

Suggests may not have been just you, routers unfortunately don't always ship with the latest firmware, so its worth trawling Linksys site to see what updates and fixes are resolved

In reply to a post by MrSaffron:

http://community.linksys.com/t5/Wireless-Routers/WRT...

Suggests may not have been just you, routers unfortunately don't always ship with the latest firmware, so its worth trawling Linksys site to see what updates and fixes are resolved

Well that's a different model number, but reading the page

As for the speed, it was Windows 10 at fault once again.
It seems that Windows 10 seems to have issues when IP's change on the router, even after I release and renew everything.

A full reboot of my PC resolved the speed issue shown here (Flash Version) and here (HTML5) taking into account its the weekend when everyone is using the internet.

Sadly still no IPv6 which really was a must for me for the software I am writing.

And the latency has doubled compared to when it was on the Fritz, but I will leave it for a few days to see how it goes.

As for the BQM, there is no point setting that up yet due to I am still changing stuff on the new router and most stuff requires a reboot of it.

Paul

Ok I am done with fiddling with this router, so I have now just added the BQM to it, so it will need at least 24 hours before we see how clean it is and then another several days to see if those spikes appear.

Hopefully it will be fine.

Paul

Well I know its only been running for just under a day, but you can already see that there is a lot different in this new graph and what I had on the last router.

On the last router before it got the spikes every 6 or so hours you would see loads of red dots at the top throughout the day, I know that's about <1% packet loss, but its a packet loss.

Now on this graph for this new router there is none apart from the 3 small ones at the start but none since 3am.

Also the latency has also settled as well.

So there is definitely a good change so far, so maybe it was a firmware issue on the last router, so I am crossing my fingers for 6 to 10 days time

Paul

Or maybe just a more powerful processor with more time to spare responding to pings.

In reply to a post by RobertoS:

Or maybe just a more powerful processor with more time to spare responding to pings.

Yes, this I suspect, when handling a WAN speed of 300mbps. I wouldn't be surprised if the design spec for most routers WAN ports would be the VDSL range upto 80, maybe 100mbps. The FTTP range of 300mbps does cause many ARM CPUs a challenge.

In reply to a post by RobertoS:

Or maybe just a more powerful processor with more time to spare responding to pings.

Well I was also wondering this before I ordered it, however AVM told me via Twitter its will be fine.

Hello, the FRITZ!Box 4040 is powerful enough to handle your 330Mbps via its WAN Port. We can supply further information here: https://en.avm.de/service/fritzbox/fritzbox-4040/kno...

The WIFI-speed is like it is displayed in this table: https://en.avm.de/service/fritzbox/fritzbox-4040/kno... The speed depends on your terminal device.

I think its like was Andrew said, just a firmware issue, but either way that router might be the cause of the issue, I guess I will find out in 7 to 10 days, but it does look promising so far.

TBH, I must of got a pre-launch version due to their marketing launch was on the April 25th 2017, so its very new.

Paul

In reply to a post by jchamier:

In reply to a post by RobertoS:

Or maybe just a more powerful processor with more time to spare responding to pings.

Yes, this I suspect, when handling a WAN speed of 300mbps. I wouldn't be surprised if the design spec for most routers WAN ports would be the VDSL range upto 80, maybe 100mbps. The FTTP range of 300mbps does cause many ARM CPUs a challenge.

Well its possible, I just checked the specs, but I am either blind or the processor info isn't listed.

But I took their word that it would be fine om my connection where they replied it would be fine, even when I told them the connection speed.

Like I replied above to Rob, I have just notice it had a market launch 5 days ago and I had already had mine for a few weeks before that.

So I am assuming now that their will be some issues at first.
I know it has Wi-Fi issues where it just drops connections now and then, and that packet loss etc

Also I have also found out its not really meant for the UK, I was just lucky Amazon had one in stock from a seller.

It was just strange that its fine for around 7 days then starts spiking every 6 hours around the time it first connected.

But this new Linksys WRT 3200 ACM-UK has a 1.8 GHz DUAL-CORE CENTRAL PROCESSING UNIT (CPU) according to the specs.

*** update ***
You might be right, I just looked in the latest firmware and I can see a mention of a Dual ARM Cortex A9, not sure if that's what's in there.

I could open it up and see if need be, I probably won't be sending it back and if it comes down to it, I could always run it as a Wireless AP.

*** update 2 ***
Just opened it up and its using a IPQ4018 which is a Quad Core ARM Cortex A7 running at 717MHz.

[IPQ4018 SoC]
Wave-2 802.11ac SoC for Routers, Gateways and Access Points

Specs here more info here (PDF)


Paul

Edited by PaulKirby (Mon 01-May-17 00:13:54)

In reply to a post by PaulKirby:

Well its possible, I just checked the specs, but I am either blind or the processor info isn't listed.

I'll have a look at that PDF, but the normal issue is that running 300mbps through NAT via the CPU can be the problem without hardware assistance. You may have luck looking in one of the gigabit ISP forums (B4RN, HyperOptic etc) as those customers have issues using most retail routers.

Its normally the NAT that overloads the CPU - its unclear if you need an Intel Core i3/i5 or similar as you find in enterprise Cisco/Juniper kit, or any of the ARM CPUs originally designed for low power and now mostly in phones.

Number of cores generally doesn't help, as routing / NATting is a single threaded activity.

The CPU spec information tells you what it can do, but doesn't tell you how fast it can route or NAT from WAN to LAN or vice versa.

On the smallnetbuilder site, there is a report that the 3200 ACM can do 500mbps WAN to LAN throughput (I assume hardware assisted NAT) at a max speed of 521 mbps.
https://www.smallnetbuilder.com/tools/charts/router/...

Edited by jchamier (Mon 01-May-17 10:53:35)

Well as far as I can see IPQ4018 is a SOC (System on Chip) with all the router stuff (like 5 port Gigabit Ethernet Switch, Dual Band Wireless, Hardware NAT engine, plus more) embedded inside, also included is the Quad Core ARM Cortex A7 running at 717MHz.

I don't think its a hardware issue, but more of a firmware issue.

Paul

In reply to a post by PaulKirby:

Well as far as I can see IPQ4018 is a SOC (System on Chip) with all the router stuff (like 5 port Gigabit Ethernet Switch, Dual Band Wireless, Hardware NAT engine, plus more) embedded inside, also included is the Quad Core ARM Cortex A7 running at 717MHz.

Yes, they all are, have been since the original Linksys BEFSR41 which I had on my NTL cable modem. (imported from US). However the BEFSR41 could not cope with any faster than 20mbps WAN speed.

I don't think its a hardware issue, but more of a firmware issue.

If you mean firmware in the Qualcomm SoC then maybe, but if you mean firmware as in the Linksys (aka Buffalo) decisions, then also maybe.

My old ASUS RT-N66U was tested with a WAN to LAN via IPv4 NAT at 500mbps with hardware support enabled. However if the hardware support was disabled (e.g. you used advanced features that needed to look at packets, such as QoS or filtering) then the WAN to LAN data rate dropped to 150mbps maximum. This was tested by small net builder using test rig on each side of the device.

PPP adds overhead that requires the CPU to encapsulate each packet, which is one reason some ISPs now don't use it (the biggest being notably Sky) as their networks do not have to host other ISPs. BTwholesale has no choice. The "MER" is just a login based around authenticated DHCP.

I was referring firmware issue to the IPQ4018 SoC that is in the last router (i.e. FRITZ!Box 4040) the one that had all the packet loss spikes ever 6 hours after several days of fine connectivity.

I have read loads of people having issues with the Linksys WRT 3200 ACM router, but it seems fine to me so far apart from IPv6 not working.

But that is probably down to having the wrong setting set for it and I was waiting for the full 7 to 10 days to verify not packet loss was happening on this router first.

There was some faffing about on this Linksys router to get working, whereas the Fritz 4040 apart from the username, it all worked out of the box.

Paul

Ok, I was just told via twitter that the FritzBox 4040 has a throughput from WAN to LAN of between 900Mbit to 940Mbit and sent me this link:
https://en.avm.de/service/fritzbox/fritzbox-4040/kno...

But I am very happy with my Linksys WRT 3200 ACM, not long setup DDNS so minimal down time when BT changes the Dynamic IP.

Paul

WAN to LAN throughputs can sometimes be impacted on any extra encapsulation used, e.g. PPPoE adds overhead in terms of CPU handling, where as most router manufacturers quote simple IP throughput

In reply to a post by MrSaffron:

WAN to LAN throughputs can sometimes be impacted on any extra encapsulation used, e.g. PPPoE adds overhead in terms of CPU handling, where as most router manufacturers quote simple IP throughput

Yeah, but going by their worst case speed of 900Mbit should still be fine for 310Mbit (down) and 31Mbit (up) combined etc, but after 10 days if you recall I was getting huge packet lost spikes that increased every 6 hours until the router was rebooted.

I am not getting it yet on this Linksys WRT 3200 ACM, granted I have only been monitoring for a couple of days, but this router has been running end of April and BT really love to re-issue those dynamic ip's, sadly BT don't do a Business FTTP product and only do a FTTC Business fibre option, well apart from the leased line, but that's out of my price range LOL.

So there is no chance in getting a static IP anytime soon from BT, I know I did have a nice long chat about this with BT this morning, stating its a shame Static IP's are not available as an add-on feature for home connections, not all would need or want it, but for those that do could just pay the extra £6 p/m.

I was told that "might" and I say "might" be an option later on, the same with FTTP (Infinity 4 equiv speeds) business package later on, its just not atm due to the lack of FTTP rolled so far.

But my DDNS I have setup (not fully automatic yet) so I will expect to see a 100% red bar sometime this or next week , just need to see how long it takes for the DNS update with a TTL of 60 seconds will take to propagate out.

But I know the initial adding of the A record only took 2 to 3 mins for Googles DNS and for your BQM to update, so I assume an IP update should be the same

Paul