Speed issue with new fiber connection

[Chomiq]

ISP guys connected fiber at my flat today. The way it's setup:
Comes in with fiber -> Huawei HG8310M -> cat 5e (0.5m) to WNDR4300v2 Netgear router -> Cat 5e (2m) to PC

Connection is rated at (up to) 600 Mbps.

Best I could get from running speedtest is ca. 270 Mbps down, 170 Mbps up on what is pretty much:

I know it's up to but I'm wondering if the old netgear router can be limiting the speed somehow?
Link to specs https://routerchart.com/netgear/netgear-n750-wndr4300-v2-wndr4300v2-289

WAN/LAN is connection status is listed as 1Gbps in router and in Windows.

Any ideas? I can get in touch with ISP to verify speeds with their supplied Huawei AC router and also try to switch 5e cables to Cat 6 for router/PC side of the connection.

 

[Cranky5150]

I experienced that same thing mate ! when i went to my half gig connection i had the same issue. I was running a old gigabit router myself. I purchased a much newer one, and i got all the speed i was paying for. Also, try updating you adapter drivers in windows as well.

 

[TheLostSwede]

Time to get a new router. You got a very old SoC in that thing, even when it was new, it was not really cutting edge.

Netgear WNDR4300v2 - TechInfoDepot

"2976482101" is silkscreened on the board in the FCC photos.

en.techinfodepot.shoutwiki.com

I guess it can't handle routing more data than what it's doing, not unusual, as the SoC wasn't built for high-speed data routing. Routing data works differently from using the built in switch for the local network, so even though it has Gigabit switching abilities, it doesn't mean it can route the data at the same speed.
Try plugging the computer straight into the Huawei bridge, you should get the full speed doing that.

One thing that might be worth doing, is checking all the settings to make sure any CPU offload options it might have, are turned on, as well as Hardware NAT if it has it. Looking at the manual, your router seems to support neither.

https://www.downloads.netgear.com/files/GDC/WNDR4300v2/WNDR4300v2_UM_23dec2014.pdf

Whatever you get, I recommend something supported by Merlin or Voxel.

About | Asuswrt-Merlin

www.asuswrt-merlin.net

Voxel's Firmware

www.voxel-firmware.com

I guess you could try DD-WRT or OpenWRT on your router first, but they're both a bit time consuming to configure.

 

[Chomiq]

Thanks @Cranky5150 and @TheLostSwede

As suggested I've hooked up directly to Huawei bridge, bypassing the router. Speed went up to 350Mbps which pretty much matches what the tech guy from ISP said he was getting at a different customer today. So yeah, it looks like (apart from the "up to" not really matching the max) the old Netgear is the culprit.

Anyone got a decent AC router they can recommend, let's say €150 tops.

 

[TheLostSwede]

Thanks @Cranky5150 and @TheLostSwede

As suggested I've hooked up directly to Huawei bridge, bypassing the router. Speed went up to 350Mbps which pretty much matches what the tech guy from ISP said he was getting at a different customer today. So yeah, it looks like (apart from the "up to" not really matching the max) the old Netgear is the culprit.

Anyone got a decent AC router they can recommend, let's say €150 tops.

This one, with Voxel's firmware.

Netgear Nighthawk X4S (R7800-100PES / R7800-100UKS) | Preisvergleich Geizhals EU

✔ Preisvergleich für Netgear Nighthawk X4S (R7800-100PES / R7800-100UKS) ✔ Bewertungen ✔ Produktinfo ⇒ Modem: N/A • WAN: 1x 1000Base-T • LAN: 4x 1000Base-T • Wireless: WLAN 802.11a/b/g/n/ac (Wi-Fi 5), simulta… ✔ WLAN-Router ✔ Testberichte ✔ Günstig kaufen

geizhals.eu

Really reliable hardware in that one. Haven't had to touch my router in years, besides firmware updates. Never had a more reliable router and that comes from someone that used to work at a router manufacturer. Not that we made crap, but the hardware wasn't as good. Admittedly it was also before the hardware inside the R7800 was available so...

Yes, the R7800 is "old" but Netgear is still supporting it and 802.11ax is still very much "beta" testing.
The most recent official firmware came out on the very last day of 2019 and added support for more Wi-Fi channels.

https://kb.netgear.com/000061548/R7800-Firmware-Version-1-0-2-68

 

[Bill_Bright]

I've hooked up directly to Huawei bridge, bypassing the router. Speed went up to 350Mbps

It's important to remember speeds can vary quite a bit minute to minute - and especially throughout the day. I would go back an forth between direct connect and through the router a few times checking your speeds. You could also try different switch ports on your router - and for sure, a different cable.

I'm not trying to talk you out of getting a new router. Since that one only supports 11n and USB 2.0, I personally would want to upgrade eventually. But if running a few more tests with your Ethernet connection reveals the times are closer when going through the router - you may not need to upgrade with the same sense of urgency.

 

[Cranky5150]

A Nighthawk is a excellent choice for a really reliable router. I use a Dlink personally and have had zero issues with it. It's been the most consistant router i have ever used TBH.

 

[TheLostSwede]

A Nighthawk is a excellent choice for a really reliable router. I use a Dlink personally and have had zero issues with it. It's been the most consistant router i have ever used TBH.

Which model and more importantly, are they following the FCC ruling now and releasing firmware updates? Asus does on some models, but the ones they don't sell in the US, gets one or two updates and are then "forgotten" about...
Personally I can't recommend D-Link due to their poor software support and the fact that I know they are more of a branding company than anything else. They don't develop hardware or software in-house and they're obviously not a manufacturer any more, as everything is outsourced. Not the kind of company you want to trust with something as important as networking hardware.
Sometimes it's a huge downside to have worked in an industry, as you know way too much of what's going on behind the scenes...

 

[Cranky5150]

Which model and more importantly, are they following the FCC ruling now and releasing firmware updates? Asus does on some models, but the ones they don't sell in the US, gets one or two updates and are then "forgotten" about...
Personally I can't recommend D-Link due to their poor software support and the fact that I know they are more of a branding company than anything else. They don't develop hardware or software in-house and they're obviously not a manufacturer any more, as everything is outsourced. Not the kind of company you want to trust with something as important as networking hardware.
Sometimes it's a huge downside to have worked in an industry, as you know way too much of what's going on behind the scenes...

I hear you sir. I used to use Netgear all of the time and i still like their products and recommend them. This Dlink just works for me though LOL

 

[TheLostSwede]

I hear you sir. I used to use Netgear all of the time and i still like their products and recommend them. This Dlink just works for me though LOL

But that doesn't answer my questions

 

[Cranky5150]

Sorry friend LOL......I've had a few......

 

[Chomiq]

It's important to remember speeds can vary quite a bit minute to minute - and especially throughout the day. I would go back an forth between direct connect and through the router a few times checking your speeds. You could also try different switch ports on your router - and for sure, a different cable.

I'm not trying to talk you out of getting a new router. Since that one only supports 11n and USB 2.0, I personally would want to upgrade eventually. But if running a few more tests with your Ethernet connection reveals the times are closer when going through the router - you may not need to upgrade with the same sense of urgency.

You're right it's a difference between 360 through the bridge (who knows this thing might have a crappy Huawei SoC in it) and 260-270 through router (tested over 4 h yesterday). Upload tops out at 200.
Thing is R7800 would cost about €200 and at the moment it makes no sense to me to upgrade if the cost itself would cover 2 years of internet bill.
Thanks for all the input guys.

 

[TheLostSwede]

You're right it's a difference between 360 through the bridge (who knows this thing might have a crappy Huawei SoC in it) and 260-270 through router (tested over 4 h yesterday). Upload tops out at 200.
Thing is R7800 would cost about €200 and at the moment it makes no sense to me to upgrade if the cost itself would cover 2 years of internet bill.
Thanks for all the input guys.

Did you not see the link I sent? It's just over €131 in Poland. Direct link to the shop below.

Netgear Nighthawk X4S R7800 AC2600 Gaming Router - Wireless router Wi-Fi 5

1 036,00 zł Router bezprzewodowy, przełącznik 4-portowy, GigE, Wi-Fi 5, Dwuzakresowy. Szybka dostwa

www.proshop.pl

And the Fibre to Ethernet bridge is NOT the weak link here, don't use that as an excuse. It doesn't route data, it's simply an optical to electrical converter.
I also thought you said they provided you with a Huawei router, at least give it a try and see what speeds you're getting.
On top of that, you should call your ISP, as if you're paying for up to 600Mbps and only get 360Mbps with your PC directly connected, something is wrong. Never make excuses on behalf of a service provider, if they don't deliver what you pay for.

I pay for 250/100 at my parents flat here in Sweden, and I get a little more in terms of upload speed and about that in download speed (keep in mind it's early afternoon now, so not the ideal time to test).

 

[Chomiq]

Did you not see the link I sent? It's just over €131 in Poland. Direct link to the shop below.

Netgear Nighthawk X4S R7800 AC2600 Gaming Router - Wireless router Wi-Fi 5

1 036,00 zł Router bezprzewodowy, przełącznik 4-portowy, GigE, Wi-Fi 5, Dwuzakresowy. Szybka dostwa

www.proshop.pl

And the Fibre to Ethernet bridge is NOT the weak link here, don't use that as an excuse. It doesn't route data, it's simply an optical to electrical converter.
I also thought you said they provided you with a Huawei router, at least give it a try and see what speeds you're getting.
On top of that, you should call your ISP, as if you're paying for up to 600Mbps and only get 360Mbps with your PC directly connected, something is wrong.
I pay for 250/100 at my parents flat here in Sweden, and I get a little more in terms of upload speed and about that in download speed (keep in mind it's early afternoon now, so not the ideal time to test).

Proshop is a Danish retailer, their Polish branch is pretty much a phone number and email address. Regular retailers have it listed at ca. 945 PLN.
ISP can supply Huawei router, I'll contact them to test it and give them the result from plugging in directly to the bridge.
For comparison monthly cost is 65 PLN vs 55 PLN for 600 and 300 respectively. So if they won't be able to deliver 600 on a clean connection I'll get them to drop the price to 55.

 

[TheLostSwede]

Proshop is a Danish retailer, their Polish branch is pretty much a phone number and email address. Regular retailers have it listed at ca. 945 PLN.
ISP can supply Huawei router, I'll contact them to test it and give them the result from plugging in directly to the bridge.

Sorry, but I don't see the problem if they have the product in stock at a much better price. This is part of the pro's with the EU.

I presume the router they provide is a loaner, that doesn't cost anything? I wouldn't trust using a Huawei router for shit when it comes to security updates though.

 

[Chomiq]

I presume the router they provide is a loaner, that doesn't cost anything? I wouldn't trust using a Huawei router for shit when it comes to security updates though.

Nope they charge you for it a one time fee. Neither would I that's why I wanted to stick with the old one.

Called the ISP they'll test it at my place next week.

 

[Bill_Bright]

Sadly, loaners still cost consumers. The "rental fee" may not show up as a separate line item on the bill, but it's still in there. And as a loaner, it can mean you are making payments on it years after it has been payed off.

BTW, if the loaner is old and outdated, you can often call the ISP up and request a trade-in for a newer model at no additional cost to you. Worth a phone call to find out. I bought my own modem and wireless router but know others with the same ISP (and other ISPs too) who have done that. And I did with it with the cable/DVR box for my cable TV.

 

[TheLostSwede]

Well, this is an ok alternative otherwise. Hopefully that's a "local" local shop.
It's another model that gets regular firmware updates from the manufacturer and Merlin supports it. Getting a bit old, but still a solid choice.

Router ASUS RT-AC68U najlepsza cena, opinie - sklep online Neonet

Router ASUS RT-AC68U najniższa cena sklep internetowy Neonet dane techniczne opinie Router ASUS RT-AC68U opinia testy recenzje darmowy transport szybka dostawa promocja cenowa

www.neonet.pl

 

[agentnathan009]

Also, you need to keep in mind that Speedtest.net isn't the most reliable indicator or bandwidth. I have a friend who works in IT and he could fill your ear about better ways to determine bandwidth. Anyway, with speedtest, the equipment on the other end may have traffic, or it may not be able to handle that level of bandwidth. For example, maybe their equipment has access to 300Mbps internet pipeline and your bandwidth can reach up to 500 Mbps. You won't see 500Mbps testing to that server and when you factor other traffic and inefficiencies you won't see the full 500Mbps of your up to speed anyway. I have found servers that could handle higher bandwidth than others on speedtest so you are getting a number that doesn't mean a lot because you could connect to a server to download a game to you game console but due to traffic and connection bandwidth available for that server you won't see the higher bandwidth.

Basically, you can have higher bandwidth than many servers and since, like with public wifi, you will get throttled so others can utilize a portion of the available bandwidth the server has access to. Say 400Mbps, if you are only person connected, but say 14 people are connected and downloading, that 400Mbps gets divided up so that comes out to 28 Mbps each, factor in inefficiencies and you might only be able to download around 20 Mbps.

 

[TheLostSwede]

@agentnathan009 fair points, although we don't really know what the OP tested against.

A lot of people recommend https://fast.com/ these days, but it also depends on where you live. I get really odd numbers testing against fast.com here.

My ISP where I live provides their own, custom Speedtest server, although I have slower internet there, since they charge too much for the next step up...

Even so, you have to agree that 1. his current router isn't fast enough and 2. there's something seemingly off if he gets half the bandwidth he's paying up to for, no?

 

[Chomiq]

What I'm testing against is a speedtest.net server hosted by the same company that's supplying the fiber connection that's used by my ISP, so this should at least show the quality of connection on the ISP->fiber provider route. Server is located in a major city 25 km away:

10 minutes later, pretty much same result in regards to DL speed:

Steam, which usually can top out the connection during download, was showing 30,6MB/s yesterday, which transaltes to roughly 240Mbps. This all through the old Netgear, plugging in directly through the bridge would give something around 360Mbps on speedtest to the same server (at least when I tested at 11PM last night) but PPPoE wouldn't hold for long in Windows 10.

Once the ISP comes here to test the connection (which is next week) with the Huawei router I'll know if the situation can improve dramatically with a router upgrade. If it does reach 500+ I'll consider going for something better than a Huawei unit that costs 100 PLN (ca. $25) to buy (not loan) from the ISP (so I doubt it's something amazing). I've called the ISP today and when I told the guy that connection through the bridge topped out at 370Mbps his response was that going directly through the bridge to onboard might not be enough to hold higher speeds.

If even with Huawei unit the connection will not be able to keep above 400Mbps of the "up to 600Mbps" that's on the contract I'll tell 'em to drop me to the 300Mbps plan. There won't be a problem and I don't get charged until next month.

I'd rather avoid throwing money on a new router just to find out that it won't do much because ISP's infrastructure is the culprit here.

Anyway, we'll find out next week.

 

[TheLostSwede]

That Fast.com result is terrible imho.
Below is this 250/100Mbps connection and it's possible to go all the way to 1000/1000Mbps here, for a price I'm not willing to pay.

 

[Bill_Bright]

you need to keep in mind that Speedtest.net isn't the most reliable indicator or bandwidth. I have a friend who works in IT and he could fill your ear about better ways to determine bandwidth.

Then what is the most reliable indicator of bandwidth? Its hardly helpful if you don't tell us all those better ways to determine a more accurate bandwidth your friend told you. Many of us here at TPU work in IT and surely would like to know.

For example, maybe their equipment has access to 300Mbps internet

fair points

No they are not fair points because that's a totally unrealistic "what if" arbitrary hypothetical scenario! For one, it is just wrong to fault Speedtest.net for "maybe" accessing 300Mbps equipment! It actually makes no sense for any testing service to do that. "Maybe" your friends method has access to only 200Mbps internet! "Maybe" Speedtest.net has access to 2,000Mbps Internet. Since that IS available in some areas, it would make sense they would have access to that. With speeds up to 10,000 Mbps (for business networks) in some areas, it would make no sense Speedtest.net (or any other testing service) would connect to such a slow poke server as a 300Mbps without - at least not without offering other options. So they don't!

If you note the Speedtest FAQ, they make it clear there are many variables that affect the scores, including the fact different servers will perform differently - that's why they let you choose from over 10,000 servers around the world!!! And they recommend performing several tests with several different servers to get the most accurate picture of your speeds. Typically, you can even test with your own ISP and that's important because then you can make a solid argument your ISP can't refute.

Plus, you can always compare your Speedtest.net results with Fast.com or Testmy.net or several other sites - any of which could be affected by your same totally unrealistic "what if" arbitrary scenario!

FTW, I get the following,
Speedtest: 146.21Mbps download and 10.41Mbps upload
Fast.com: 150Mbps download and 8.7Mbps upload
Testmy.net: 138.8Mbps download and 10.9Mbps upload

Which one is right? I suspect all of them were - at the time they tested and for the server they connected to. But if you Google what is the best way to measure Internet bandwidth? Speedtest is most often mentioned.

Note that many home routers let you test from within the router's admin menu. This is probably one of the best ways for home consumers because it eliminates any latency issues caused by your computer or its NIC.

 

[Chomiq]

For comparison:
Speedtest to fiber provider - 270 Mbps down 170 Up
Fast.com - 130 Mbps, 140 Up
Testmy - 10 Mbps (this stores results based on IP so it has archival results from March this year when someone run it from the same ISP, topped out at 300 Mbps), 65 Up

So yeah quality going outside is all over the place.

 

[agentnathan009]

Then what is the most reliable indicator of bandwidth? Its hardly helpful if you don't tell us all those better ways to determine a more accurate bandwidth your friend told you. Many of us here at TPU work in IT and surely would like to know.



No they are not fair points because that's a totally unrealistic "what if" arbitrary hypothetical scenario! For one, it is just wrong to fault Speedtest.net for "maybe" accessing 300Mbps equipment! It actually makes no sense for any testing service to do that. "Maybe" your friends method has access to only 200Mbps internet! "Maybe" Speedtest.net has access to 2,000Mbps Internet. Since that IS available in some areas, it would make sense they would have access to that. With speeds up to 10,000 Mbps (for business networks) in some areas, it would make no sense Speedtest.net (or any other testing service) would connect to such a slow poke server as a 300Mbps without - at least not without offering other options. So they don't!

If you note the Speedtest FAQ, they make it clear there are many variables that affect the scores, including the fact different servers will perform differently - that's why they let you choose from over 10,000 servers around the world!!! And they recommend performing several tests with several different servers to get the most accurate picture of your speeds. Typically, you can even test with your own ISP and that's important because then you can make a solid argument your ISP can't refute.

Plus, you can always compare your Speedtest.net results with Fast.com or Testmy.net or several other sites - any of which could be affected by your same totally unrealistic "what if" arbitrary scenario!

FTW, I get the following,
Speedtest: 146.21Mbps download and 10.41Mbps upload
Fast.com: 150Mbps download and 8.7Mbps upload
Testmy.net: 138.8Mbps download and 10.9Mbps upload

Which one is right? I suspect all of them were - at the time they tested and for the server they connected to. But if you Google what is the best way to measure Internet bandwidth? Speedtest is most often mentioned.

Note that many home routers let you test from within the router's admin menu. This is probably one of the best ways for home consumers because it eliminates any latency issues caused by your computer or its NIC.

First of all, I'm not in IT, so trying to regurgitate testing with a Linux box and certain other hardware/software to get accurate bandwidth measurements gets over my head. He rattled off some things that he did for internal testing, and that conversation took place last year, so I'm sorry if I can't give you more details other than his information sounded quite compelling as I listened. Conversation started with 5G bandwidth being promised by cellular carriers and field tests with 5G wide band capable phone tests.

Second, I did make valid points because those are variable that the average Joe doesn't think about. I gave examples to help illustrate that "up to" is just that, up to and you have to take into account numerous other variables and bandwidth sharing with other people. There are various servers and they may not support the upload as fast as he can download. Take Comcast for example, they offer 25Mbps download and a wimpy 5Mbps upload, at least they did offer such speeds when I was using their service. Verizon offers 100Mps up and down. Another variable that the average Joe won't take into account. They see a speedtest number and complain if it doesn't match what the ISP says they should get. I explained some variables that OP did not take into account with his testing, since most were pointing him to upgrade his equipment. I don't have equipment that I can bring over that I know is capable of 1Gbps to replace his to test and see if it is his equipment or other variables that could be affecting his bandwidth results.

Testing to different servers on speedtest can help you see, as you put at the bottom of your post, that there are variables, but your bandwidth is so low you are less likely to see the possible limitations of servers.

Next time, how about you use more constructive dialogue rather than bashing what someone is saying. Some of us are attempting to help the OP sort out his issues rather than attacking each other for perceived misinformation...

For comparison:
Speedtest to fiber provider - 270 Mbps down 170 Up
Fast.com - 130 Mbps, 140 Up
Testmy - 10 Mbps (this stores results based on IP so it has archival results from March this year when someone run it from the same ISP, topped out at 300 Mbps), 65 Up

So yeah quality going outside is all over the place.

I would hope that your ISP isn't over promising on speed and under delivering. I would be interested in the results the ISP gets testing your equipment.