Jump to content

rebrecs

Members
  • Content Count

    35
  • Joined

  • Last visited

  • Speed Test

    My Results
  • Days Won

    4

Reputation Activity

  1. Thanks
    rebrecs reacted to Dave Taht in Bufferbloat! (latency under load)   
    Both the dslreports folk and fast.com reached out the the bloat email list (see lists.bufferbloat.net) as to how to go about measuring this problem properly in their codebases. You will find a lot of good info in the archives there, and we're always looking for sites to be actively testing for bufferbloat. Of the two, dslreports has thus far been doing a great job, so great that their dataset is thoroughly polluted by people that used the site to fix their bufferbloat!!, so we no longer have a real picture of what the internet is really looking like. (so I really, really, really applaud the idea of a new site, such as yours, attempting to tackle the problem also) 
     
    I have a few nits on the dslreports stuff I've always wanted them to address, also. A few are:
     
    0) huge threads on the bloat lists that I won't summarize... a noted one is the insistence on doing some level of statistical ledgerdemain on the data (throwing out the worst 5% of the data, or picking an arbitrary threshold of X latency for bufferbloat, etc. )When it comes to this sort of science, the *really* interesting data is in the outliers, not the averages. 
     
    This is a detailed look at that sort of statistical rigor problem from a talk I gave at sigcomm 2014: http://conferences.sigcomm.org/sigcomm/2014/doc/slides/137.pdf
    (they've never invited me back)
     
    1) Since the adoption of fq_codel in OSX, openwrt, thousands of commercial routers (notably now in Wifi - see google's implementation here: http://flent-newark.bufferbloat.net/~d/Airtime based queue limit for FQ_CoDel in wireless interface.pdf ) and the universal enablement of ECN in that OS, we are starting to see ECN negotation and CE markings show up in multiple data sets. It would be good to track that somewhere. 
     
    2) both dslreports and fast.com throw out too much data. The really core and scary bufferbloat problem is when a network is too congested to operate worth a dang in the first place. I keep hoping that someday dslreports, at least, will create a plot that just shows the data they currently throw out - an analogy of what we might discover is here: https://www.space.com/25945-cosmic-microwave-background-discovery-50th-anniversary.html
     
    3) I really like the http://www.dslreports.com/speedtest/results/bufferbloat?up=1 plot - my kvetch is that it is only a 10 day most recent summary and I've had to rely on screen shots to be able to compare stuff over time. I'd long hoped for a deal where they could sell or share that dataset to researchers. The bufferbloat problem IS getting better - assuming the dslreports dataset isn't totally polluted but there is a long, long way to go.
     
    4) Nobody's tests run long enough to saturate higher speed links, due to how slow TCP ramps up. A variable length test, or one that runs longer when it detects high bandwidth is in use. dslreports cuts off their data set and test with 4+ second delays - and we have seen delays as bad as hundreds of seconds in the field.
     
    5) A really simple test would be to measure syn and syn/ack times while under load for a string of very short tcp transactions. This would emulate web traffic better.
     
    6) Recently published (and under discussion on the bloat list) was a pretty good summary of the speedtest problems we have on the internet going forward. Discussion here; https://lists.bufferbloat.net/pipermail/bloat/2019-May/009211.html - the paper, here: 
    https://arxiv.org/pdf/1905.02334.pdf  
    Anyway, we're kind of old internet fogies that mostly use email, and not web forums like this, if you have further questions, want to gain testers, or have someone from the bufferbloat effort or academia help dissect the data, please drop us a line on bloat at lists.bufferbloat.net.
     
    Best of luck with it! Thx!
  2. Like
    rebrecs got a reaction from Pgoodwin1 in checkout this amazing result   
    Wow... look at this test result. I pay for 400 Down and my Avg Down is around 448. I snuck one by their speed throttle mechanism. At least now I know my neighborhood and facility wiring etc. can handle 1 gig. How this happened I don't know, but here is my theory. Based on the last thing I did, which was manipulating my so called router to grab it's Ip address from the host's DHCP but ignore everything else that comes from the Host's DHCP -- my router finally picked up DNS from 1.1.1.1
    I proved my so called router was using the new DNS server via the cool tool on this site (DNS lookup or something similar, under Misc --> tools).
    So, the very first time I ran the test without the extra DHCP baggage - my Down test hit 1 gig+, but their clever infrastructure caught on real fast. At first it over-corrected, then settled down in at the speed I usually get in my test results. So, no, benchmarking is not my hobby. I just want to be well prepared for my inevitable conversation with my dear old ISP.
     

  3. Like
    rebrecs got a reaction from CA3LE in checkout this amazing result   
    Wow... look at this test result. I pay for 400 Down and my Avg Down is around 448. I snuck one by their speed throttle mechanism. At least now I know my neighborhood and facility wiring etc. can handle 1 gig. How this happened I don't know, but here is my theory. Based on the last thing I did, which was manipulating my so called router to grab it's Ip address from the host's DHCP but ignore everything else that comes from the Host's DHCP -- my router finally picked up DNS from 1.1.1.1
    I proved my so called router was using the new DNS server via the cool tool on this site (DNS lookup or something similar, under Misc --> tools).
    So, the very first time I ran the test without the extra DHCP baggage - my Down test hit 1 gig+, but their clever infrastructure caught on real fast. At first it over-corrected, then settled down in at the speed I usually get in my test results. So, no, benchmarking is not my hobby. I just want to be well prepared for my inevitable conversation with my dear old ISP.
     

  4. Thanks
    rebrecs got a reaction from CA3LE in ISP (Host) average   
    I can easily accept that rationale. Thank you.
    I don't want to get started on what they "deserve."
     
    I continue to evangelize the use of TMN
    Good stuff.
     
  5. Thanks
    rebrecs got a reaction from CA3LE in A couple of router settings that matter   
    Sir Webmaster,
    I am using DNS services from Cloudfare (1.1.1.1 / 1.0.0.1)
    So that you know your efforts are not in vain - I chose Cloudfare based on reading "Step 6" a while back. Good Document.
    😉
     
     
     
  6. Like
    rebrecs got a reaction from Sean in ISP (Host) average   
    Hi, this is just a small thing .. but since you asked....
    After a test, there is comparison data at top. Things like % > than my average, % greater than whole world, etc.
    One of them is % > than Host. Hmmm.
    I would have to think about this longer to know what to recommend. For now I will just submit what I am pondering.
     
    The average numbers on Host have a lot to do with the subscription. (e.g. I pay for 5Mbps, and so 5Mbps is what I get.)
    Lots of people are running tests, and we don't know what speed they are paying for in their subscription.
    I think it may drive assumptions in peoples minds that could give the ISP a bad rap. As in, it may not be true that XYZ ISP can only go 5 Mbps Up, but it is certainly true that the majority of people here running tests have paid for that specific rate.
     
    Like I said, its a small thing, but does mean that particular comparison statement might could use an asterisk or something to clarify the fact that the Host results are sensitive to more than just the capability of their infrastructure. It is not an issue for individuals running tests since we already know the specifics of our own subscriptions.
     
  7. Like
    rebrecs got a reaction from CA3LE in ISP (Host) average   
    Hi, this is just a small thing .. but since you asked....
    After a test, there is comparison data at top. Things like % > than my average, % greater than whole world, etc.
    One of them is % > than Host. Hmmm.
    I would have to think about this longer to know what to recommend. For now I will just submit what I am pondering.
     
    The average numbers on Host have a lot to do with the subscription. (e.g. I pay for 5Mbps, and so 5Mbps is what I get.)
    Lots of people are running tests, and we don't know what speed they are paying for in their subscription.
    I think it may drive assumptions in peoples minds that could give the ISP a bad rap. As in, it may not be true that XYZ ISP can only go 5 Mbps Up, but it is certainly true that the majority of people here running tests have paid for that specific rate.
     
    Like I said, its a small thing, but does mean that particular comparison statement might could use an asterisk or something to clarify the fact that the Host results are sensitive to more than just the capability of their infrastructure. It is not an issue for individuals running tests since we already know the specifics of our own subscriptions.
     
  8. Thanks
    rebrecs got a reaction from CA3LE in A couple of router settings that matter   
    Hi all, just passing along my experience - your mileage may vary.
    I think its worth submitting since I was able to double both my UP and DOWN test speeds with these small changes.
    These observations were made with an ASUS AC-3100 Router, from a hardwired 8p 24 gig (mem) server with a 1 gig Ethernet card.
     
    Firewall off
    I understand this is religious, but I set the devices to protect themselves as best I can. After all, I had no firewall rules in place anyway. So the firewall was just kind of a big piece of code in the router that had to do work on the packets - slowing the router down significantly. IMO that protection can be done elsewhere for much less expense in speed. Again, broad stroke firewall (even with no rules set) does protect against certain attacks - so turning it of is a choice I gamble on since I'm not the department of defense and nobody is trying to prove anything by screwing me over.
    AND, my server firewall does the same stuff regarding DoS attacks and what not. ( all these fear factors we live with )
    NAT Acceleration (cut through) on
    NAT is a confusing thing to investigate. Devices, L3 switches, routers all have settings. I'm still reading about it. However for today, I can tell you that NAT acceleration ON allows a bunch of stuff to bypass the processor in the router and it makes a HUGE difference in Download speed. Turning it on or off did not make any difference in Upload speed.
    DNS Settings
    Talking about DNS settings arrives at a discussion of DHCP settings. That is because DHCP does you the kind favor of loading up the address of it's favorite DNS values when you do not tell it otherwise.
    This is the setting that allowed me to get from about 5Mbps Upload to 20Mbps Upload instantly. I was initially using the ISPs DNS server, and had my router set to allow DHCP pass through so my devices were becoming contaminated with the ISPs DNS address. This is a tricky topic but it makes a big difference.  Consult your manuals carefully.
    Bottom line, do what is needed to get a good DNS server working for you.
     
    --john
     
     
  9. Like
    rebrecs got a reaction from Sean in A couple of router settings that matter   
    Hi all, just passing along my experience - your mileage may vary.
    I think its worth submitting since I was able to double both my UP and DOWN test speeds with these small changes.
    These observations were made with an ASUS AC-3100 Router, from a hardwired 8p 24 gig (mem) server with a 1 gig Ethernet card.
     
    Firewall off
    I understand this is religious, but I set the devices to protect themselves as best I can. After all, I had no firewall rules in place anyway. So the firewall was just kind of a big piece of code in the router that had to do work on the packets - slowing the router down significantly. IMO that protection can be done elsewhere for much less expense in speed. Again, broad stroke firewall (even with no rules set) does protect against certain attacks - so turning it of is a choice I gamble on since I'm not the department of defense and nobody is trying to prove anything by screwing me over.
    AND, my server firewall does the same stuff regarding DoS attacks and what not. ( all these fear factors we live with )
    NAT Acceleration (cut through) on
    NAT is a confusing thing to investigate. Devices, L3 switches, routers all have settings. I'm still reading about it. However for today, I can tell you that NAT acceleration ON allows a bunch of stuff to bypass the processor in the router and it makes a HUGE difference in Download speed. Turning it on or off did not make any difference in Upload speed.
    DNS Settings
    Talking about DNS settings arrives at a discussion of DHCP settings. That is because DHCP does you the kind favor of loading up the address of it's favorite DNS values when you do not tell it otherwise.
    This is the setting that allowed me to get from about 5Mbps Upload to 20Mbps Upload instantly. I was initially using the ISPs DNS server, and had my router set to allow DHCP pass through so my devices were becoming contaminated with the ISPs DNS address. This is a tricky topic but it makes a big difference.  Consult your manuals carefully.
    Bottom line, do what is needed to get a good DNS server working for you.
     
    --john
     
     
  10. Like
    rebrecs got a reaction from CA3LE in Results Question Please - what is this number?   
    Hey CA3LE, I looked up combined average. I actually like that formula. I just wasn't aware of it <anymore>. As you said, hard to say what is most meaningful for ranking. I suppose ranking Up and Down separately has a couple of things going for it. The ISPs sell their product in terms of two speeds, and have pretty much taught the public that Up and Down are different. 
    The combined average, while elegantly capturing the weight of each individual average set, is not pure with respect to network speed. (i.e. includes a weight factor for # of events in each average set, and measures the test operators choices of which test) If some are interested in characterizing the test operator's decisions, they could look at the raw data and find out how many of each were run.
    Along those lines, I was recently considering how the ISP data limits, or even the Cost of the subscription might be useful -- then regained my senses. Corrupts the central theme of Net Speed.
    Maybe someday, when you SELL the Pro version, you can make it a comprehensive benchmark :-)
    It is hugely useful and wonderful just the way it is.
    So that's my vote FWIW. Separate Ranks. I have not done any screen programming since the Jurassic period, but I imagine you are all over it !!
  11. Like
    rebrecs got a reaction from CA3LE in Results Question Please - what is this number?   
    This screenshot is from "Member Statistics" results page.
    How is this number computed please?

  12. Like
    rebrecs got a reaction from CA3LE in TMN text and expressions: does 'host' = 'provider' ?   
    I think it's probably fine. Especially if I am the 1st and only to ever ask. Admittedly, I am not the best at associating a meaning via context. I liked what you said; "Host = ISP because they are hosting my internet connection." Given that I was operating a web test on a web test site, the simplest explanation would be the one you intended. Though I will say, my second guessing arose from the possibility that TNM may have captured my domain and used it to test the connection speeds to the Host that Hosts my websites or the  Host that Hosts my email, etc. However, reading further I now Know there are a fixed set of participating servers. (hosts)
    "Server" is another one of those words.
     
    As a sidebar I will tell you that the section on improving connection speed was good. I found a couple of not nice things on my Windows box after reading that.
    And, as I read through the section on 1.1.1.1 DNS - I was forced to learn a lot more about DHCP than I ever wanted to know. I was hosting DHCP on my router for the LAN but all it was doing was using DHCP pass through from the ISP so I was still getting their DNS, and so forth and so on.
     
    Anyway - great site. I love this test. I run it every morning.
     
  13. Like
    rebrecs got a reaction from CA3LE in TMN text and expressions: does 'host' = 'provider' ?   
    I'm sure some will find this nit picky. Sorry, I have to ask. Pasted from a test result, see the word Host on the button. That word is highly overloaded. I just want to confirm my assumption that it means ISP infrastructure. When I press that button, it is not immediately apparent to me what I'm looking at exactly. I am still hunting for the help page that describes it.

  14. Like
    rebrecs got a reaction from LisaGraham in multithreaded testing   
    Can anyone suggest a doc to read on this topic? I am familiar with multiThreading in a general programming context. I understand why networking Apps might benefit from multiple connections.
    But how does our TNM test use it?
    I did a search on this forum for "multiThread" and found a few posts, all of which were oriented toward someone trying to achieve a specific result with responses typically saying "try multithread."
    OK. Well, "try it you will like it" is ok I suppose.
     
    If anyone is aware of a tutorial document that might help illuminate why and under what circumstances a multithread test is appropriate, please hit me up.
     
    My Environment contains Windows 10 clients, MAC OS clients, cell phones, and TVs connected to the internet. My hope is to determine which type of testing is most meaningful. I am not looking for the "good news" test (i.e. I don't run marketing benchmarks).
    Currently I run a test per day and watch the pattern (more Mbps or less Mbps) without a clue as to why it might be going up or down.
     
    Thanks in advance -- to anyone caring to educate the Newbie.
     
    --john
     
     
  15. Like
    rebrecs got a reaction from mudmanc4 in New User Hello   
    Hi all. Sometimes you just have to have more cowbell ! I can't explain it.
    I'm on the edge of the Houston metropolitan area. Mostly farms a few years ago. Civilization is doing what it does; ever expanding outward.
    That means, (you guessed it) Pizza delivery and internet !
    I have all the same grief everyone else does. My ISP has given me more than enough reasons to keep an eye on them.
    Additionally, I am doing research into hosting my own domain.
    By way of introduction, that is me in a nutshell. I'm looking forward to hearing about what others are experiencing.
    The testing App? So far so good, though it is certainly not the "good news" app. (ha ha)
    Sincerely,
    --Rebrecs
     
     
×
×
  • Create New...