New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
i think you are talking about yourself. when i tried to get something technical from you, to understand why your script was failing, and suggest a feature to the script, all you showed was your toxicity and complete lack of concern for why the script was failing. https://lowendtalk.com/discussion/comment/4440319/#Comment_4440319
you turned down someone who was trying to help you for free in the worst possible way, all your comments are invalid to me by default after that. i don't want to argue more, this is just for others to see
You did not, as is clearly visible in the thread you linked, "try to get something technical from me" - whatever "something technical" may mean - but you confronted me with assertions and no evidence that I did anything wrong.
You also did not try to help me, but you rather suggested that I somehow include a shell script from you in my testing, which btw. was double senseless because (a) why would someone include a shell script from an unknown person in a compiled program?, and (b) because I saw that your script used iperf3, which (1) is just a marketing tool, and (2) there already are plenty of iperf3 based tools out there and plenty of (IMO quite useless) "benchmark" results from it.
If you really just had asked questions I would have responded constructively, as I already did many times when people had (true) questions.
Funny side note: quite often during times when I was attacked or my benchmark called bad I actually helped a provider with that very benchmark program to check a new node or new node type or new config (like e.g. a new disk array and testing different Raid configs) upon their request.
So, whom and whose judgement should I take more seriously, a provider's who has put lots of money on the table (for their new nodes or routers or ...) -or- some guy on a forum who, like in your case, wants me to use his script and if I don't personally attacks me?
But OK, you wanted another hit (attempt) at me and you had it. Can we now return to the topic of this thread?
asking how your test works and why it failed multiple times is technical and constructive enough for me, you never answered
your benchmark is not comprehensive enough and sometimes fails for unknown reasons, i tried to fix this and make your benchmark better, but you refuse to do anything with it with silly reasons.
"consider adding this" doesn't mean copy and paste my code, it means analyze what it does, how it works, why it could be useful, and then look if you can replicate with your own code.
so real-time monitoring and detailed reporting is marketing? wow, just wow. so your own test, which is just a "download me a file from x, no realtime monitoring/reporing, literally nothing except a speed value and ping time in the end" (and it's not clear if it's single or multi-threaded) which sometimes fails for unknown reasons, is somehow better and not a marketing tool?... i think the answer should be clear for everyone.
im not saying iperf3 is perfect, IMO it's average (reporting could be better and there aren't enough features for a comprehensive test), i just don't know a better tool.
that's good, but your network test is just a "wget bigfile.bin" (i assume bc you refuse to explain), if a provider can't do it by themselves, that's a bad sign
constructive criticism is not an attack, there is more i would want to say, but you are sadly a brick wall, are autistic by chance?
yes, please, im also tired of your nonsense. no one is perfect, and i hope to see good quality comments from you and better benchmarks in the future.
I did explain that multiple times. You could have looked for it, but you chose to just ask and presume that I had to answer for the umptieth time. Well, you were wrong.
Besides some unknown user on a forum willy nilly saying that my benchmark failed doesn't mean that it's true, nor that I'm somehow forced to prove otherwise.
Because some unknown user on a forum says so? And btw repeating some unfounded willy nilly BS statement doesn't make it true.
Moreover, there is proof that you're just spreading BS -> simply look at the version number, it clearly indicates that I did improve functionality and took care of initial bugs.
Stop your BS already! I did look at your script, I did analyse it - how else would I know that it's using iperf3?
For a start I didn't say that my benchmark program is better than iperf3. What I said is that my focus is different and that my tool serves its purpose while iperf3 doesn't (and basically lies).
Mainly though, sorry, but it seems that you really don't know what you're talking about. What I do is not just "basically wget me a file", nor is there "no realtime monitoring/reporting". In fact I originally wanted to just use some decent http client library, maybe libcurl but I found that some functionality I wanted and needed wasn't there and adding it in would have been more work than writing my own code. Just one example: "download 1 slice of say 1 MB and then pause for x ms", because you see, I mainly benchmark VPSs that is systems on a node with shared resources! So I obviously don't want my benchmark behave like a reckless pig. But of course that also necessitated complex time measurements with microsecond precision, which allows my benchmark to on one hand measure, compute, and show the real speed while on the other hand being a "good neighbour" on a node and not grabbing all the resources for myself.
BS!
Btw, you seem to not yet have noticed it but I'm not testing dream machines but real VPS, almost always exactly the ones customers get - which means that resources are limited. Plus, when I wrote my benchmark (1st version) very often one didn't get TBs of traffic volume. So, when benchmarking multiple days just doing "wget bigfile.bin" wasn't an option. I had to make do with a compromise (typ. 100 MB) or else - and that did happen a couple of times! - my bandwidth would be limited due to using too much bandwith/traffic.
Oh, now you're escalating to really nasty personal attacks?
Well, just ignore my reviews then. Nobody forces you and you anyway already took the liberty to call my work and my (quite patient btw) responses "nonsense". Yeah, a truly classical case of "I just have some questions, plus I want to help".
Help someone else!
Maybe now we can back to the topic of this thread?
That's because it is. The actual fiber route, which @TimboJones referred to, has a direct impact on latency.
Other factors are involved that may make things slower, but you're not getting around the physics. The shortest fiber route from one location to another will always tell you the minimum latency possible.
Yes, correct - well, to a large degree because fibers rarely run the direct, straight, shortest route. But still, yes, the actual route as well as geo distance are major factors.
OTOH, I quite often see significant differences from one location to another, say Denver - WDC between different providers. Short version: routing is a major factor as well. For instance I more often than rarely see e.g. 7 hops vs. 12 hops or, especially with american carriers outright insane routes like e.g. HE mercilessly and completely needlessly inserting a very significant detour via for example (and typical) Seattle.
Marfa, Marfa Texas is the correct answer.
Remember hops mean nothing as a lot of providers hide these which can make them not show all of the hops using mpls.
Example if Seattle to NY shows 5 hops from one provider and 12 from another doesn't mean the 12 is worse.
Same, I also wasn't thinking of India as Asia and was only thinking of China, Japan, etc. Fair enough.
My bad. I've said all this over the years, just not as detailed on the linked comment.
Where I talk about post secondary education in Electronics Engineering (I know you don't accept that, they taught about IPv6 in the year 2000!) and then a career as a QA manager. I did not mention it was for a wireless radio manufacturing company that made rugged, outdoor radios for which I have had HANDS ON experience with networking and throughput testing on a daily basis pretty much close to 13 years. I should have also mentioned testing with hardware traffic testers, as well as both proprietary testing software from wireless chipset makers and with open source tools like iperf. And there's many, many Fortune 500 companies using iperf for repeatable and cross manufacturing testing (despite having bugs), how many people are using your app?
And then my current position as Hardware Qualification Engineer, not with networking, however, but points to background in test methods and requirements.
But if only you could provide an example of your qualifications that would be "tangible", that would have saved a lot of time. You asked me my fucking qualifications (I think 10,000 hours experience is legally accepted as an expert) and I did just that. If you want a copy of my diploma, you can, I'll just stick it right up your ass. That'll feel real tangible.
You keep talking about providers using your app, but don't realize your numbers are bogus (untrustworthy) and they talk shit about you in DM's. So I wouldn't use that as your "tangible qualifications".
I'll give you the benefit of doubt and just believe that.
As for the rest, re myself, in particular
there's an interview on LEB telling quite a bit about where I come from and my background.
Btw, testing software vs designing and developing software are quite different things.
As for the DMs that's not surprising; those who don't like me contact someone who very much dislikes me, and those who do like and/or trust me and my competence contact me. But of course there's a decisive difference: those who ask me for testing or advice or finding the cause of some troublesome or weird phaenomenon have money/investments on the table, those who contact you, oh well, trash talking is cheap...
Yep, I know, but more often than not the 12 hop route has higher latency than the 5 or 7 hop route.
This is especially true over short distances.
Netcup handles this very badly lol. Even tho their DCs are inhouse, they have a 4-5 hops to backbone/Internet
>
Wow, so generous. What an asshole remark. Say that to a person to their face and see what reaction they give.
Jesus Christ, what a waste of time, you could have replied in two sentences. The fucking nerve to demand my qualifications when yours are severely lacking. Your experience doesn't support your ego.
I'm even more annoyed you majored in physics at University and still said "geography doesn't necessarily mean a whole lot wrt the internet" where your criteria is good bandwidth and many connections.
No shit. There's a reason QA testers exists. Programmers make bugs. I've tested your app, provided test results and offered servers for test to reproduce the issue but you stuck your head in the sand.
I can point to half a dozen developers who would praise my bug finding skills. Not a single one reacts like you, like you made perfect code impervious to faults. Can you point to several teams of developers who rely on you to QA their work?
Companies where devs do their own QA are garbage, as you've demonstrated.
That doesn't hold up since they contacted both of us (me after you). They can also come to the conclusion you're full of shit AFTER engaging with you and your feedback.
Pay me for my services. They are also much larger than these LET hosts.
Indeed, as I have exactly zero reason to trust you. But it's not just generosity, the main factor is that I don't perceive talking with you as any fruitful.
I also once worked for a carrier and so I know that whatever city, or more likely town if not village, happens to be in the geographical center of the USA is very unlikely to have huge bandwidth and excellent connectivity. In other words, you should have listened to what @raindog308 explained ("backhaul").
As for the rest of what you wrote I do not even care to respond.
Have a nice day.
This never ending, and will happen again.
That's NOT what you wrote, as it would have made sense. You changed your tune.
Again, backhaul means all your traffic goes through ONE link. So that is nonsense when asking for well connected.
Anyway, enough of this shit.