Jump to content


  • Content Count

  • Joined

  • Last visited

  • Speed Test

    My Results
  • Days Won


Posts posted by rebrecs

  1. On 3/8/2020 at 12:39 AM, CA3LE said:

    Reality is that overhead really depends on the size of the transfer.  It's probably in the neighborhood of 3-7%... more likely on the lower end.  Meaning that 1.73 - 1.8 Gbps would be needed instead of 1.68 Gbps in the scenario above.  So you might lose 50 to 120 Mbps in overhead on such a fast connection but I feel that's pretty minimal in the grand scheme.  If you're pushing nearly 2 Gbps... do you really care about losing 0.05 to 0.12 Gbps of it?  ... I think I could manage.  :-P 

    Hi, thanks for the links! I will definitely have a look. Also, I understand your unspoken guidance. (like, get out of the weeds dude :-)) All I really have to do is reveal that I thought about it and took a cursory look. I won't let it become a lifestyle. I promise.

  2. On 3/8/2020 at 7:46 AM, mudmanc4 said:

     Packet size effects throughput as well, due to the number of packets required. As well as payload size. Is n truly equatable outside of a proofing environment?

    Thank you!

    The only "knob" i can find to turn on "packet size"  in the switch gear is "allow/ don't allow Jumbo Packets."

    I have not found any "settings" to influence packet size in the TV interfaces.

    I suspect there are knobs to turn on the boxes with full blown Operating Systems but they are not really the devices that concern me at the moment.


  3. Hi,

    Looking for a good reference book on networking but,

    There are lots of networking books for sale, and mountains of research papers - all of which concentrate on some part of networking. I am not sure what part of networking might hold the answers to the questions I have. I'm hoping to limit my search by first asking here. Looking for - anything published that looks at bits and bytes in terms of 'work?' Perhaps some characterization of work versus overhead? I checked my test data csv file and our conversion here at TMN is pretty much a straight "divide by 8," which makes perfect sense; a byte is 8 bits all day long. Got it.

    I'm just going to ramble for a paragraph or so, and maybe someone will make sense of what I think I am asking;

    How much of the bit-stream on the wire constitutes a byte of actual data in the way a customer would think of their data? I believe that is a question about overhead. In my view, (as a customer) if I am going to move a 200 MB file, then in the simplest terms, and using networking byte representation,  [200,000,000 bytes x 8 = 1,600,000,000 bits]. Without asking any more questions - I go away with the understanding that it takes 1.6 Gbits to move my 200 MBytes. But is that true ? (ignoring the silliness that networking byte math introduced)

    For discussion, and with no basis whatsoever for choosing the numbers, if in reality it takes 2 Gbits to move my 200 MBytes, then the overhead is introduced by what? I am going to guess it would come from the protocols, handshakes, CRCs, fields left blank -- I don't know, whatever protocols do.

    Since any bits or bytes that are purely overhead still consume bit cells, then they use time. So, do they get measured in a Mbps measurement? Or discarded/factored out?

    How do it all work?

    So that is the play pen I am going to jump into for a little while.  In case there are any recommendations, thank you in advance.

    As well as recommendations to reference material, anyone "in-the-know" who can lay it out for me - please by all means - let me hold it. 



  4. I'm interested in this. And, no need to apologize.

    As far as one party screwing up what what the other one does, that is what they do. The people who elect them count on them to do that.

    I would love to talk about personal freedom and human rights but at the moment I really would like to hear about the open internet.

    I am actually writing something about that, so please - how is the open internet being taken away? (Sincere question)

    I can definitely use some ammo.





  5. On 1/3/2020 at 9:47 PM, spenceteeth said:

    I run speed tests out of boredom I have seen all kinds of weird things across national and international tests. I wouldn't be concerned

    Thanks SpenceTeeth. And I noticed CA3LE hit the "LIKE" button so I'm assuming ca3LE agrees with you. But dude(s), "pay no attention to the man behind the curtain?" (Wizard of Oz) Really?

    23-3/4 hours per day I totally agree with "not to worry." The devices on this local net are all running well. But for the 15 minutes per day that I'm thinking about this stuff-- the result is either 1) erroneous, or 2) it reveals an upload limitation beyond what I currently know about.

    I suppose you could say, that's what I do "out of boredom." :-) But, that would not quite be true.

    On 1/3/2020 at 9:47 PM, spenceteeth said:

    Possible your ISP and test server for Dallas are one in the same.

    That also raises several interesting questions. Beginning with, "what is the speed toll (if any) for packets handed off from one ISP to another?" 

  6. As part of today's testing, I ventured out into new territory by changing the mirror location. Until today, all single path testing has been done with the Dallas mirror server.

    After selecting all the various US based mirrors, an unexpected and perplexing story unfolded.


    I get the same speed test results using any mirror in the US on download speed tests.


    On Upload speed tests, I get my ISP capped limit (40Mbps) from Houston to Dallas, and I get approximately 15Mbps consistently from Houston to any other mirror in the US.

    I have not tried the mirrors outside of the US.


    So, what's going on with that? There is the obvious factor-- latency, but why would latency only affect uploads?

    For completeness, I will ask the following questions as well; 1) "could I be operating the tests incorrectly?" 2) Is there something special about the Dallas mirror server? 3) Has this been seen before?


    Every day is a fun day ! Thanks in advance to anyone able to assist.


  7. On 12/5/2019 at 8:07 AM, rebrecs said:

    Suddenlink fell from 1st to 2nd place with a noticeable reduction in average speed.


    Their infrastructure in my geography got very sick earlier this year and it took them a long time to straighten it out. Thus it makes me nervous when I see this drop.


    Having more experience now with TestMy, I realize the significance of the ISP is not as I had believed earlier. The ISP rank is formed by the tests the TestMy members (and guests) are running. My connection to suddenlink, for example, (aka suddenrage) is pretty fast so I can drive the Suddenlink rank up by running a bunch of download tests. I can watch it change because it updates very quickly, unlike the member rank tab on the suddenlink page which updates, seemingly, only by special request. Again, the ISP rank is a result of whatever people are doing at the moment. Thus the results should be treated that way.

  8. On 12/27/2019 at 10:53 AM, mudmanc4 said:

    @rebrecs, Have you entertained the idea of using a 1:1 NAT for the website(s) using port forwarding for various web servers ?

    Hi @mudmanc4,

    I would have to read a couple of books or go take a course just to understand what you just asked me. So, no.

    Time permitting, I may.

    I have often read manuals where Port Forwarding settings are identified. They tell me what do do, were I to desire to forward a port. However, I have yet to find anything that tells me why I would want to do such a thing, or what the typical (or intended) use of it is.


    Just curious, sidebar: off topic, what do you have on your head in that picture?

  9. On 12/20/2019 at 6:39 AM, Sean said:

    The best way to run this type of test would be to run iPerf in server mode on one PC and then run simultaneous iPerf speed tests against the iPerf server from the other devices.  You can download iPerf here.  To use it, open two command prompts and run "iperf3 -s -p 5201" in one and "iperf3 -s -p 5202" in the second.  To run simultaneous tests on the other PCs against it, run "iperf3 -c -p 5201 -t 60" on one PC and "iperf3 -c -p 5202 - t 60" on the second PC. 


    If there's a spare LAN port on the router, plug the PC running iPerf in server mode into it.


    iPerf shows the current bandwidth every second, so you can monitor how the bandwidth is split over the 60 second test duration.

    I don't know. I'm not certain that is going to reveal what I'm looking for, but it could still be interesting. I think it depends on what the so called router does with the traffic. If all of the traffic is sent to the ISP's routers to determine what to do with it, then the test result may be of interest. All of the hardwired ports on that box are bridged. It has one port named the modem port which is an actual gateway port. It should send local traffic to the local bridge but that is not a guarantee. I'm no routing expert, but the 192 addresses are assigned the local bridge interface, but the 129 address gateway is which points to a suddenlink server. Which one wins ? I have been in situations before when suddenlink was down and things on the local net could not see each other. Other times they could. I might pitch this thing and get a real router. Who knows.


    Kernel IP routing table
    Destination     Gateway         Genmask         Flags   MSS Window  irtt Iface UH        0 0          0 eth0   U         0 0          0 br0   U         0 0          0 eth0       U         0 0          0 lo         UG        0 0          0 eth0


    Ultimately, my purpose for testing, and the original reason I got involved with TestMy was/is in preparation for self-hosting one of my domains. My facility is combined residence/business. I pay the ISP the residential non-commercial rate. The domains are currently hosted at BlueHost.

    After reading about NAT, it seems that it may cause issues for my self hosted domain, Rather than try to itemize all the ways NAT would interfere, I decided to that getting on IPv6 would be a good idea. My ISP, which is Suddenlink/Altice, is silent about IPv6 on their site. I am (time permitting, and it takes a LOT of time) attempting to contact a human being at Suddenlink to ask them about it. Some time ago I asked them about getting a static IPv4 address from them - they won't do it. When I am able to ask my question regarding IPv6 support, I anticipate their response to be - we don't have it.

    If they cannot support IPv6 connections,

    I could get an IPv4 address some other way, and just start using it. (whether they like it or not)

    What would their NAT machinery do if a packet came in with an IPv4 address  outside of the 192, 10 , 100 etc range of NAT-able addresses? It would have to honor it (forward it) .. right ?




  11. I spent a lot of time getting Upload tests to run at the full ISP speed. Until I figured a out correction that got rid of the mid-var, my Uploads had terrible speed (5Mbps). This caused me to associate mid-var with bad speed. But look at this download test result graph. This time the mid-var works in my favor. I am referring to the upward bump in the middle.This test was run on a connection limited to 400 Mbps (ISP tiered cap). Although it is preceded by a downward bump, the downward bump never went under 400.image.thumb.png.ffa4132e9c4d186ced237eb6f8261356.png

  12. On 12/10/2019 at 5:44 PM, rebrecs said:

    my guess would be they are throttling at the signalling level versus measuring and capping a "byte flow" inside their switching gear. If I am right about that (and that's a big if) then the modem at their end, that your modem talks to, would set the max transmission speed.

    I may need to revise that remark. Lately I am seeing instances of data points on my results graphs that are much faster than the ISP speed limit for my account. What tends to happen it is a bit of oscillation above and below the speed limit, finally  converging on a more or less steady set of data points, and indicating a lot of middle variance. I have not changed anything. This is new. But, it the speed limit was set by setting fixed modem clocks, data would not be able to go faster than the modem speed. Thus, I remain open to the idea the ISP is measuring something else. (queue length/time, or packets / time <?>)


    I think I will remain open also to the notion that this could be a side effect of the test, or the browsers, or the time-stamping on the systems under test.

  13. 11 hours ago, Sean said:

    Whenever you use cellular data, public Wi-Fi, privacy VPN, etc. your public IP address is likely shared with several hundred other users.  So if one runs TestMy on one of these connections, there is a good chance that someone else could run TestMy at the same time on the same network,  I.e. it should have no problem running simultaneous tests from the same IP address, but with one test potentially skewing the other.

    Sean, thanks, I really wish I understood what you have said here. Are you saying an ISP (via their DHCP server) will issue the same IP address to many different customers ?

  14. Background

    My ISP speed is clamped at 400 Mbps Down, though I have been routinely getting 430+ Mbps.

    I have a 1 gig infrastructure in house. The direct-attach router ports are consumed by L2 switches and one Wintel computer.  I execute TMN tests (Loc 1=multithread, Loc 2= normal) from the Wintel box.


    Reconfigure for new test plan

    I am about to introduce two more locations (TMN Identifiers -Loc's 3/4) which will be a pair of laptops. When attached to available L2 switch ports, each of the laptops produce TMN download test results near identical to the Wintel box I have historically used. I intend to use the laptop pair to run concurrent TMN download tests. My goal is to see how the speed allocation is arbitrated. I believe theoretically the answer should be near 50/50, or approximately 200 Mbps results for each device.


    The Questions

    Questions are centered around how TMN is going to react to this.

    • Will TMN test controller (for lack of a better name) allow this?
    • If so, Will TMN think this is one test or two ? If two, will the Identifiers be preserved?
      • Each laptop browser will be initializing a test using different Identifiers. Will that cause two different Test IDs even though they share an IP address?
      • What happens if I push the "go" button very near the same instant? Would it be better to delay the start of the second system by a second or two?

    Please advise on how I should proceed.


    Thanks in advance



  15. Interesting. So, if their own "blessed test" is a meg off, I think you already have a case to present. I don't know whether you already have <?>

    Are you mandated by CenturyLink to use that specific router? I am really not happy to suggest spending money, who the heck wants to do that. But, an ISP only believes one configuration is legitimate when discussing your issues - and that is for your systems to be attached directly to a modem. In your case, you cannot do that since it is all built into one integrated entity. If you were inclined to separate the modem from the switching and Wifi gear, someday, it might be in your interest to do so.

    The ISP techs, the kind that come to your house, have a gadget they use to hook to the modem. I would be interested (if it were me) to see what reading that thing produced.

    Those gadgets produce all sorts of results, including REAL jitter (not network jitter) based on the clocks in the data stream. They also produce results regarding signal level (usually in db) and they produce results in Mbps !! The thing we want. If you don't have a discrete modem, then having a tech run the gadget is a good idea.


    In your original post you asked whether your modem could be slowing you down. Maybe. But I think more than likely, if the culprit lies anywhere in the a90-7500, my bet would go on the switch/router settings. I say that because if the modem part of the box were getting transmission errors, they would see it at their end too, and your log would be filling up.

    So, service call for tech+gadget, and check all the logfiles.

    That's my $0.02



  16. In this context, jitter is a statement of deviation. In fact, engineers/techs who only work on tcp/ip networks will also use the word synonymous with "late packet arrival."

    I have never looked at what exactly the n% means but lower is better and I think 2% is probably ok.I saw something about it on Wikipedia, but It gives me a headache to think about it simply due to my objection to the word being used in this context at all. In all other areas of engineering it describes a deviation from a fixed expected rate. I would love to hear somebody tell me how a packet switching network has or ever claimed to have a fixed expected rate. but anyway, that's just me. It likely means 2% of the time, packets arrived later significantly later than the mean. (taking aspirins and guessing)


    I still think chasing down that middle variance may get you closer to your 6Mbps. Especially since both tests are about 1Mbps off. If you fix it, I bet your upload will jump up too. I'm only saying that because it made such a difference for me.

    There are people on here that legitimately know a lot more about this than I do. I'm hoping one of them in particular will come to the rescue.

    What kind of router do you use ?


  17. RMcQ, yes, TestMy.Net is a great resource. I never see any ads so no telling how they get paid. Possibly selling the test results. Or maybe collecting from the ISPs for a better service resource than they have :-) Who knows.

    Anyway, back to the topic, does centurylink have a speed test of their own ? If so, it is likely an Ookla based test. The results of those have to be used carefully, but they can provide an interesting data point. I ask, because you have a 5% middle variance. I spent some time fighting middle variance and it was the ISP Ookla speed test that gave me the hints I needed to finally fix it. When my TMN tests were indicating a high middle variance the ISP test was also bursty. The ISP speed test always returned perfect numbers (i.e. the full paid-for speed) but on its meter I could see the traffic stopping ans starting in bursts, which agreed with the middle variance result on my TMN speed test. My variance may have been greater than 5, and the notes are not handy. I don't know how bad the variance has to be before the bursts can be seen on Ookla. But, i WOULD LIKE TO

  18. Interesting question. There are several ways to do it. Each has less or more merit depending on the goal. Given that an ISP's goal is the  managing speed limits on a per customer basis, and the fact they tend to discuss (and sell) "speed" in terms of bits-per-second, my guess would be they are throttling at the signalling level versus measuring and capping a "byte flow" inside their switching gear. If I am right about that (and that's a big if) then the modem at their end, that your modem talks to, would set the max transmission speed.

    I really don't think it would be practical for them to do otherwise.

    That's my $0.02



  19. Hi, sorry for your troubles. I can totally relate to what you have been through.

    Just curious, if you don't mind ... the gear that they brought to your house, is it mandatory to use their equipment? Asked another way, do you have a choice of using your own internet access equipment? (modems, routers, switches , whatever)


    Second, the Hughesnet equipment, what are the brands/models they use ?

  • Create New...