Geekzone: technology news, blogs, forums
Guest
Welcome Guest.
You haven't logged in yet. If you don't have an account you can register now.


View this topic in a long page with up to 500 replies per page Create new topic
1 | 2 
  #3175538 26-Dec-2023 12:42
Send private message

Are you sure on that? A gigabit for 8 seconds (fairly typical speedtest length) is moving a gigabyte of data. Doing that 12x per hour is 12GB/hour, or 288GB/day. On its own, that probably puts you in the top couple of percent and is certainly something that would be untenable if everyone did it. 

 

Chorus says average usage is more like 600GB per month


 
 
 

Shop Mighty Ape for electronics, games, computers books and more (affiliate link).
Lias
5351 posts

Uber Geek

ID Verified
Trusted
Lifetime subscriber

  #3175539 26-Dec-2023 12:46
Send private message

SomeoneSomewhere:

 

Networks are never built with the assumption that all users are using 100% capacity 24/7. There are massive diversity factors inherent in the design - for starters, Chorus has 16x potentially-gigabit customers per upstream port, sharing a 2.5Gb/s capacity. It only gets tighter from there. Look at the HFC networks and it's a few hundred customers sharing similar bandwidth. 

 

Why do you need to prove whether or not you're getting headline speeds 24/7?

 

 

I'm well aware of the sort of contention ratios inherent in most broadband deployments and the pitiful CIR's on residential UFB connections. But equally I know that neither of those are generally an issue in NZ (on UFB at least) and thanks to SamKnows/TrueNet we have some evidence of that.

 

You could ask the same question of anyone who's participated in the SamKnows/TrueNet monitoring programs, but for me it's a bunch of things. I like data and SamKnows is useful, but I like having my own, independent data and I want that data to be as granular as possible. Having 5 minute granular data in this instance let me provide evidence that the OP does have a leg to stand on, whereas if I was only testing hourly the chances are it wouldn't show the same level of volatility.





I'm a geek, a gamer, a dad and an IT Professional. I have a full rack home lab, size 15 feet, an epic beard and Asperger's. I'm a bit of a Cypherpunk, who believes information wants to be free and the Net interprets censorship as damage and routes around it.


Lias
5351 posts

Uber Geek

ID Verified
Trusted
Lifetime subscriber

  #3175541 26-Dec-2023 13:06
Send private message

SomeoneSomewhere:

 

Are you sure on that? A gigabit for 8 seconds (fairly typical speedtest length) is moving a gigabyte of data. Doing that 12x per hour is 12GB/hour, or 288GB/day. On its own, that probably puts you in the top couple of percent and is certainly something that would be untenable if everyone did it. 

 

Chorus says average usage is more like 600GB per month

 

 

I'm probably wrong, but if for argument's sake we accept your figures, 288GB/day is still only using less than 3% of the theoretical max of the connection, it's hardly excessive.





I'm a geek, a gamer, a dad and an IT Professional. I have a full rack home lab, size 15 feet, an epic beard and Asperger's. I'm a bit of a Cypherpunk, who believes information wants to be free and the Net interprets censorship as damage and routes around it.




michaelmurfy
meow
12507 posts

Uber Geek

Moderator
ID Verified
Trusted
Lifetime subscriber

  #3175542 26-Dec-2023 13:06
Send private message

I just checked. Each one of my tests on a gigabit connection uses around 1.5 Gigabytes. You can also see the JSON payload if you go into any of the tests in speedtest tracker. 

 

Every 5mins = 18gb every hour, 430gb daily. Yes, it’s not cool doing this every 5mins. Hourly is fine. 





Michael Murphy | https://murfy.nz
Referral Links: Octopus Energy ($50 Credit) | Tesla | Quic Broadband (use R122101E7CV7Q for free setup)

Are you happy with what you get from Geekzone? Please consider supporting us by subscribing.
Opinions are my own and not the views of my employer.


  #3175544 26-Dec-2023 13:41
Send private message

Lias:

 

SomeoneSomewhere:

 

Are you sure on that? A gigabit for 8 seconds (fairly typical speedtest length) is moving a gigabyte of data. Doing that 12x per hour is 12GB/hour, or 288GB/day. On its own, that probably puts you in the top couple of percent and is certainly something that would be untenable if everyone did it. 

 

Chorus says average usage is more like 600GB per month

 

 

I'm probably wrong, but if for argument's sake we accept your figures, 288GB/day is still only using less than 3% of the theoretical max of the connection, it's hardly excessive.

 

 

 

 

That's the problem: 3% is excessive. You're using more data than 90% of the country just for statistics.

 

Consumer connections are not sold for 24/7 100% use. Fair use policies are largely a thing of the past, but for a pretty simple reason:

 

Most people use no more total data when given a gigabit connection than a 100Mb/s connection - the peaks get higher but the averages don't really change. That's a big part of why Chorus has repeatedly ratcheted up the 'standard' speed from 20>50>100>300Mb/s. 

 

This is especially a problem with small, niche ISPs like Quic that have a) a higher proportion of 'pro' users that use lots of data, and b) a smaller pool of users, so outliers have more of an impact.


RunningMan
8107 posts

Uber Geek


  #3175574 26-Dec-2023 16:25
Send private message

Lias:

 

Even running one every 5 minutes, we're still talking a few hundred megabytes of data per hour. You cannot seriously tell me that a few hundred mb of data per hour is making any sort of noticeable impact on an ISP core network, even one with issues.

 

 

@michaelmurfy did the maths above, you're about 2 orders of magnitude out, so using vastly more than your estimate. The issue is this isn't spread evenly through the day, it's regularly peak loading at least one point in the network beyond it's capacity. Your tests quite clearly show that given the significant variation. When links get to that saturation point, packets get dropped, which is the key issue other customers are facing.

 

A typical speedtest in a single direction goes for 15 seconds. 15 seconds every 5 minutes is saturating the link for 5% of the time, which is pretty close to the 4% packet loss I've seen thrown around in some threads. Now we all know you can make numbers back up any old arguement and I am in no way saying packet loss through Quic is on your shoulders (you quite rightly point out that 1 customer is just a drop in the bucket), just trying to illustrate the point that if a bunch of customers do the same thing quite regularly, collectively it can have a greater impact than imagined.

 

This reminds me of the self fullfilling prophecy. In this case, oh, I see Quic are having a network problem so I better do a speedtest to see if I'm affected. Hmm, doesn't look that bad, I'll do another. That a looks a bit worse, I'll keep doing them. You get this positive feedback loop where all the testing traffic actually contributes to the problem rather than simply measuring it. Kind of like people stuck in a traffic jam complaining about the traffic; they're actually part of the problem - maybe they're only 1 car out of 5000, but all those vehicles combine to form a blockage.

 

 


tccki
10 posts

Wannabe Geek


  #3175585 26-Dec-2023 17:20
Send private message

Residential UFB ain’t specced to give everyone uncontended line rate @ 100% duty in 2023. Like how carriers have to deal with suboptimal routes, because physics and economics. Now that you know that, assure yourself that the infrastructure has known compromises. Your speedtest ain’t gonna tell you a thing about where an issue lies it’s just granularity as you say, there are other tools for this and some of them you can’t really use on your side. You’re paying a provider to do their job so it’s simplified to one metric: how quickly real issues are resolved like loss or if teh internets isn’t reachable.


If you want better terms you can just pay for them. Otherwise you really should reduce your speed testing to something commonly recommended like >= 1 per hour on a staggered start.



tccki
10 posts

Wannabe Geek


  #3175648 26-Dec-2023 20:13
Send private message

Fat fingered that post, <= 1 per hour

mentalinc
2854 posts

Uber Geek

Trusted

  #3175660 27-Dec-2023 08:06
Send private message

You're better off looking at the load on their network provided via the status page to understand the load.

 

As others have said, stop doing speedtests all the time, you're more likely making your internet worse as anyone who uses the internet when speedtest is running gets a poor result.

 

And unless you know nothing else is using bandwidth when you have a low speedtest result, you don't learn anything.

 

e.g. a speedtest starts at the same time as a large download, maybe an nvidia driver update or large windows update, you get a low speedtest and blame quic.. but you were just using the internet at the same time..

 

You want to be looking at a trend over time, 24 data points per day, 168 over the week will give you that trend you're looking for...

 

 

 

 

 





CPU: AMD 5900x | RAM: GSKILL Trident Z Neo RGB F4-3600C16D-32GTZNC-32-GB | MB:  Asus X570-E | GFX: EVGA FTW3 Ultra RTX 3080Ti| Monitor: LG 27GL850-B 2560x1440

 

Quic: https://account.quic.nz/refer/473833 R473833EQKIBX 


Ironsfan

13 posts

Geek


  #3177382 3-Jan-2024 18:37
Send private message

Well just to show I can be less "shouty" Sorry about the all caps but was so annoyed at the time.

 

Speeds have been a little more  consistent but the ping spikes are killing my gaming. Will give quic a little more time to sort their issues out as at this time of year very little happens with anything.

 


michaelmurfy
meow
12507 posts

Uber Geek

Moderator
ID Verified
Trusted
Lifetime subscriber

  #3177389 3-Jan-2024 19:09
Send private message

@Ironsfan as per the sticky and the other post linked in here this is something they’re working on with priority. It would have been already fixed before Christmas if it were not for some problems with their new network. They’re replacing their whole network which will sort both the speed issues along with the packet loss for all. 

 

I’m going to lock this again as it has been well explored and documented now plus Quic/Vetta are in a change freeze like everyone right now so nothing will happen. Everyone just needs to show a little patience right now.

 

Refer to this: https://www.geekzone.co.nz/forums.asp?forumid=194&topicid=311025 along with the sticky threads. 





Michael Murphy | https://murfy.nz
Referral Links: Octopus Energy ($50 Credit) | Tesla | Quic Broadband (use R122101E7CV7Q for free setup)

Are you happy with what you get from Geekzone? Please consider supporting us by subscribing.
Opinions are my own and not the views of my employer.


michaelmurfy
meow
12507 posts

Uber Geek

Moderator
ID Verified
Trusted
Lifetime subscriber

  #3178037 5-Jan-2024 14:20
Send private message

Changes posted here to address network issues: https://www.geekzone.co.nz/forums.asp?forumid=194&topicid=311313 





Michael Murphy | https://murfy.nz
Referral Links: Octopus Energy ($50 Credit) | Tesla | Quic Broadband (use R122101E7CV7Q for free setup)

Are you happy with what you get from Geekzone? Please consider supporting us by subscribing.
Opinions are my own and not the views of my employer.


1 | 2 
View this topic in a long page with up to 500 replies per page Create new topic





News and reviews »

Synology Introduces BeeStation
Posted 23-Feb-2024 14:14


New One UI 6.1 Update Brings Galaxy AI to More Galaxy Devices
Posted 23-Feb-2024 10:50


Amazon Echo Hub Available in New Zealand
Posted 23-Feb-2024 10:40


InternetNZ Releases Internet Insights 2023
Posted 20-Feb-2024 10:31


Seagate Adds 24TB IronWolf Pro Hard Drives for Multi-user Commercial and Enterprise RAID Storage Solutions
Posted 19-Feb-2024 16:54


Seagate Skyhawk AI 24TB Elevates Edge Security Capacity and Performance
Posted 9-Feb-2024 17:18


GoPro Releases Quik Desktop App for macOS and Introduces Premium+ Subscription Tier
Posted 9-Feb-2024 17:14


Ring Introduces New Ring Battery Video Doorbell Pro
Posted 9-Feb-2024 16:51


Galaxy AI Transforms the new Galaxy S24 Series
Posted 18-Jan-2024 07:00


D-Link launches AI-Powered Aquila Pro M30 Wi-Fi 6 Mesh Systems
Posted 17-Jan-2024 20:02


Newest LG 4K Lifestyle Projector Doubles as Art Objet
Posted 9-Jan-2024 15:50


More LG Smart TV Owners Set To Enjoy the Latest webOS Upgrade
Posted 9-Jan-2024 15:45


Panasonic Announces the Z95A and Z93A With Fire TV Built In
Posted 9-Jan-2024 15:30


Amazon Echo Pop Review
Posted 8-Jan-2024 14:22


Samsung Tab S9 FE Review
Posted 17-Dec-2023 08:26









Geekzone Live »

Try automatic live updates from Geekzone directly in your browser, without refreshing the page, with Geekzone Live now.



Are you subscribed to our RSS feed? You can download the latest headlines and summaries from our stories directly to your computer or smartphone by using a feed reader.







GoodSync is the easiest file sync and backup for Windows and Mac