Jump to content


Beta Tester
  • Posts

  • Joined

  • Last visited

  • Days Won

  • Speed Test

    My Results

Everything posted by FiberGuy

  1. I seed Torrent 24/7. Most of the time there is little going on, but every so often someone has a decent connection.
  2. I may not have a 1Gb connection, but I have a *dedicated 100Mb connection that I don't have to share with my neighbors and my ping is better. Almost spring! /knock on wood In other words, I'm kind of jelly. You getting that 10Gb service for $400/month? *Not truly dedicated, but the ISP provisions bandwidth such that all customers can run at 100% at the same time without overloading the GPON port. My fiber is actually dedicated all the way back to the CO. They use a flat network design, 1 hop for me to get to Level 3. Now I just need to figure out how to get testmy.net to play well with my system.
  3. It's been a while since I've last posted here. Hello! My issue seems to be that I get quite the difference between Chrome and Firefox. IE just doesn't work for the upload test, not that I care to figure out why. I know my connection is good, dedicated bandwidth to Level 3 via my ISP, but my results do not reflect that. I assume it's something to do with my browsers or OS or an interaction of the two. Here are the two patterns that I am seeing 1) Download test. Starts off strong, but then around 80%, no matter the test size, it tapers off from about 90Mb/s down to about 10Mb/s. It doesn't matter if I'm testing 50MB or 200MB test sizes, once that 80% mark is hit, it drops. 2) Upload test. Starts off weak, like sub 1Mb/s, jumps up to 10Mb/s, drops back down to 1Mb, holds for a bit, then launches up to 80-90 for the rest of the test. Average is about 35. Here's why I feel confident about my connection. This traffic graph was taken back when I had 50Mb. As you can see, I was using upwards of 80% of my connection with out my ping flinching.
  4. Here's the front of one of my new TV boxes Here's the back. Notice the RJ45 input Here's my new network in the basement. The thin cable going into the ONT at the bottom is the fiber and the thicker cable right next to it is the power coming from the UPS. The top most port on the ONT is the Internet VLAN and the second port is the TV VLAN. You can see that the installer plugged a secondary switch into the TV port, then ran two cables from that switch and punched it into the patch panel. From there, the two TV ports run directly to the set-top boxes. The extra switch is only fast Ethernet, so that means all devices share a 100mb connection to the ONT. The features that I get are I can go back and watch TV shows up to 24 hours old(yes, I can FF through commercials), so basic DVR for select channels. I get 14 basic channels, half of them are 1080p. I also get a bunch of free on-demand, which has recent popular TV series and some movies, many on 1080p. All of this for $15/month. For an extra $16/month, I can get full on DVR that allows recording of two of any channels at a time, has an easy interface to track series, even if the series is on different channels. I can even instruct it to ignore episodes that have already been recorded, even if recorded on different channels. Up to 250 hours of storage and anything recorded has no time limit, you can theoretically keep something for years. It is not locally stored, but back at the CO. There is even an android app to manage the DVR over the Internet. Since I only have the basic TV package, I don't think I will be getting the full featured DVR, but I do get a free 3 month trial. When playing around, I noticed that even though I was not using DVR on a given channel and that the channel technically was not one of those "go back 24 hours" channels, if a show was in progress, I could actually jump to the beginning of the show. So auto DVR on most channels, allowing me to watch the beginning of the current show. Ohh, that $15/month covers up to 5 of these boxes. I currently only have two.
  5. FreeBSD? Linux? I recommend FreeBSD+ZFS for a server primarily used for storage, but find out what you need it to do before you decide which Os to install. Have fun. I want a server to, need more $$$, but debt first.
  6. ROFL.. I'm a scrub... Long story short, I was playing with my TCP settings a long while back and forgot to set them back. The reason I was playing with them is I could download 50mb/s from the Internet, but transferring a file across my 1gb network was only about 1mb/s.. not bytes.. bits. Something with Window's default TCP settings was not playing nice. A quick change and I was up over 110MB/s or near 980mb/s. BTW, 33MB upload sizes does not seem to be enough for me to get a good consistent speed. It takes a second for my speed to ramp up the TCP send window, but at 4.8MB/s average, the 33MB is only a few seconds. My download is better, but still not the best What tipped me off to check my TCP settings was I found another speed test site that wasn't speedtest.net, and also claimed to use HTML5. I found it strange that my upload had the same flat packet-shaped look and almost the same speed. That was way too close to be a coincidence. SpeedTest.Net did not show me this issue... How I hate that place. It's "fun", and kind of useful, but very superficial. Looks like the CDN gives me a better download, but the upload isn't as strong
  7. Netgreat 3700 Calix ONT that works as a transparent Ethernet bridge. It's about the size of a 5 port desktop switch. The mounting was done by my ISP. If I had the TV package, they would just give me a set-top device that uses the network cables. They don't use COAX anymore, they do IPTV.
  8. Upload issue has been fixed: https://testmy.net/ipb/topic/31239-upload-inaccuracy-fix/ My upload and download speed tests are now showing about the same. Still lower than my line rate, but stable and match what my network graph shows. I love this aspect, "other" speed tests I have seen my network graph showing 20mb-30mb, and the speed result listing 60mb.
  9. 33MB upload against Dal seems to be spot on with my bandwidth graph now. Thanks
  10. I know this is a bit late, but this is the first time I have heard of a story like this. Glad to hear everything seems well now.
  11. bgp.he.net/ does not show a peers tab for my ISP. It does show it for many other ISPs. For the Berkeley test, I got all green except my computer is 5 seconds slow and my router may have some vulnerability. I really want that Netgear Nighthawk. I get almost my full 50mb/s from YouTube during peak hours, and that is not from a local CDN, but from Google in another state. As for blocked ports and other things, my ISP proudly advertises that they do not block any ports, except 25 but removable upon request, do no traffic shaping, and no QoS. All traffic is best effort, even their own IPTV and VOIP services. They do not reserve any bandwidth for their own services. Except network management protocols of course. Back when you needed to peer with or host a Netflix cache server, I asked my ISP if they had any plans to sign up to OpenConnect so I could watch in SuperHD. After a few days of escalation, I got a response from a higher-up who told me they had no plans to peer with or install a Netflix cache server. He said they have plenty of bandwidth and do not want to favor one company over another. It seems their general attack plan is to just brute force everything with more trunk bandwidth from Level 3. I'm not complaining. I talked for 45 minutes with a Senior Network tech who had been with them for over a decade. He told me they played with over-subscription and traffic shaping at one point, but it was a huge head-ache to manage properly and made unhappy customers when things went bad. So they switched over to a non-oversubscribed dedicated network. He said network management is just so much easier when you're not worrying about bandwidth. Cuts down on support calls to. I asked him about their trunk, he told me they have not had any issues with congestion for a long while. They typically upgrade well before they get near capacity of their main link. They also have a secondary and tertiary link. He said they are setup such that they can bond their primary and secondary links, and could re-route some over their tertiary. Since their internal network is all dedicated and their trunk is not congested, my only limitation of speed is with Level 3. Who am I to argue? What I would like to know is how they keep buffering latency down. My ping is about 12ms to YouTube in Chicago. If I overload my connection with uTorrent, my ping might spike to 20ms, but that's about it. After that, it's the beginning of packet loss, no buffer bloat here. I disabled QoS on my router because it's no longer an issue. The difference between 12ms and 20ms is only 8ms. I don't understand how there can only be 8ms of buffer
  12. In Wisconsin many Schools can purchase 1gb/1gb dedicated fiber connections for $300/month from a non-profit co-op. The bandwidth is purchased whole-sale from the state uni at cost. Since the state uni purchases bandwidth increments of 100gb, it's not much of an issue to get cheap bandwidth. Doing trace routes via looking glass servers, they have many trunk lines. One line going from Madison to Chicago Level 3, Another from Madison to Chicago AT&T, another from Madison to Chicago Hurricane Electric, another going from Madison to an IX in Chicago, from Madison to different IX in Chicago, another link to an IX in Madison, where AT&T and Level 3 also connect. They seem to be peering with SoftLayer at one of those Chicago IXs. Then another connection to Minnesota to Level 3, another 2 sets of connection via UW-Milwaukee down to Chicago directly to Level 3 and AT&T. All of the state Universities are connected via a private fiber network, and they all purchase bandwidth from UW-Madison. Needless to say, they don't have bandwidth issues. I would love to see some speed tests from one of those servers. I had to get some 13GB of family photos from a relative once. He tossed them up on one of his servers, told me the server has a 10gb interface, and his department has a 20gb link to the Madison trunk. My 30mb connection suddenly felt slow.
  13. In theory, no, in practice, yes. People remember bad experiences very well. If I have even one experience of congestion in the period of a month, I am going to remember it and complain about how I need more bandwidth. Rule of thumb, if you are inconvenienced for any reason because of lack of bandwidth, you don't have enough. Just wait until you have several computers on your network that all attempt to download a 1GB Team Fortress patch at the same time, while you're in the middle of <insert something important that is latency sensitive>
  14. My bill is exactly $94.94 and has not ever changed. 50/50, but the below is only me downloading. I can also get all the /29 static IPs I want for $10/month/block. I was also told that I have "dedicated" bandwidth. I ask, "dedicated like how DSL is 'dedicated'?", they replied (paraphrased from memory) "No, you don't share bandwidth to our trunk. We do not over-subscribe our network up to the trunk" I'm not expecting an enterprise SLA to back up their definition of "dedicated", but they're doing a great job. My fiber line has had 24/7 uptime with no packet loss on my line for the past 6+ months strait. I've experienced a total of about 1.5 hours of down time so far, but was caused by issues on their trunk, never my line. Boy was my wife mad. I had to remind her that this kind of downtime was the norm for the past several years, until we got this ISP. Spoiled, I tell ya. The most impressive thing I noticed was I found someone with a VPS in Los Angeles, who let me ping him. A 5,000 mile round trip and under 200 microsecond of jitter about a 1 minute of samples with 0.5seconds between samples, according to my high resolution ping, at 9pm.
  15. I have a 50/50 connection, and I seem to be getting low speeds with all speed tests. All of these tools and FAQ info seems to be worded quite well, so it seems that the makers of testmy.net know what they're talking about. Anyway, here's my tracert to the dal server Tracing route to testmy.net [] over a maximum of 30 hops: 1 <1 ms <1 ms <1 ms 2 <1 ms 1 ms 1 ms 66.x.x.1 3 2 ms 1 ms 1 ms x.x.x [x.x.x.x] 4 11 ms 12 ms 12 ms te-9-4.car4.Chicago1.Level3.net [] 5 13 ms 13 ms 11 ms ae-31-51.ebr1.Chicago1.Level3.net [] 6 13 ms 12 ms 11 ms ae-6-6.ebr1.Chicago2.Level3.net [] 7 12 ms 12 ms 11 ms ae-1-51.edge4.Chicago3.Level3.net [] 8 14 ms 12 ms 12 ms te2-5.bbr01.eq01.chi01.networklayer.com [] 9 13 ms 13 ms 13 ms ae7.bbr01.eq01.chi01.networklayer.com [] 10 34 ms 35 ms 36 ms ae20.bbr01.eq01.dal03.networklayer.com [] 11 36 ms 35 ms 35 ms ae6.dar02.sr01.dal05.networklayer.com [] 12 76 ms 40 ms 36 ms po2.fcr04.sr03.dal05.networklayer.com [] 13 35 ms 36 ms 37 ms [] Route seems quite nice. I did a speed test to the Dal server, and kept ping running during it. Ping statistics for Packets: Sent = 254, Received = 254, Lost = 0 (0% loss), Approximate round trip times in milli-seconds: Minimum = 34ms, Maximum = 38ms, Average = 35ms Looking at the numbers I was getting on my network graph, it pretty much lines up with the result, so that's good. Here's my upload Something is wrong with this result. My below graph shows my upload as green. As you can see, it was VERY stable. It was a hair over 11mb/s, but my result was 5mb/s. I noticed after my upload bandwidth dropped to 0, my test didn't finish. It sat there for as long as the upload took. https://lh3.googleusercontent.com/-OgQqcrlSxjE/Uo69emC7_MI/AAAAAAAABZA/I2TOMFneU60/w157-h223-no/TestNetUpload.png It seem it counted the time the test it sat idle as part of the calculation for throughput. So about 24 seconds of 11mb/s and another 30 seconds of 0mb/s, averaged out to about 5mb/s, my final result. Here's another photo of my network graph of me downloading a file from NewYork City around 9pm, many months ago. https://lh6.googleusercontent.com/-8jeh-Z2kQpM/UZV1EcR6KzI/AAAAAAAAA98/88GxCu3M6NU/w159-h230-no/DownloadGraph.png Here's a photo of my network graph when I tried to load test. I was downloading from Steam around 40mb/s, then loaded up uTorrent and tried seeding about 100 Linux ISOs. This was also around 9pm. https://lh6.googleusercontent.com/33nNmK6Wz2kvpwVFGDpTNKOa3rNxEVV9eVJwhAXRS00=w440-h207-p-no Here's the results of running ping against YouTube for the past 4 days. Btw, my ISP does not peer with anyone or has any caching servers. All data is over their trunk. Ping statistics for Packets: Sent = 337658, Received = 337415, Lost = 243 (0% loss), Approximate round trip times in milli-seconds: Minimum = 10ms, Maximum = 26ms, Average = 11ms I'm curious as to why my up and down speed tests are so low. Thanks
  16. Using file transfers and looking at my NIC bandwidth, I get better speeds to these places than the speed test servers. Japan, Taiwan, China, India, Australia, Singapore, South Africa, North-East Africa, Arabia, South America, Central America. It seems this site needs a Chicago presence. I can appreciate the effort put into making a good speed test, but the "TestMy.net Bandwidth Test Legitimacy" section of the site makes it sound like these tests are nearly infallible and if there is a speed issue, it's your ISP. In my case, it is not my ISP, but the hosting provider which this service is running on. There are always corner cases, but it's a bit haughty to say "Our servers are also configured and tested to maintain full quality of service for thousands of miles." when they bottle-neck in some corner cases is the internal network of the Speed Test service. It's not just this site. I get similar issues with SpeedTest.Net and a few other sites. They seem to connect with Level 3 in the local region, then route their own inter-regional traffic. While my speed tests tend to be low, my bandwidth works where it counts. I get great speeds to real services. Kind of an inverse situation from my last ISP, where I got great speed tests and relatively poor real world speeds.
  17. It seems the servers being used are from SL(SoftLayer). My ISP's upstream provider is L3(Level3). SL has 80Gb/s of peering with L3 in Dallas, but I am in the Midwest. It seems SL only has 10Gb of peering with L3 in Chicago. What happening is that my tracert goes from my ISP, to L3 Chicago, to SL Chicago, to SL Dallas. Now, SoftLayer only has a total of 20Gb of their own private bandwidth inter-connecting their datacenters. When I tracert SL Dallas, the jump from SL Chic to SL Dal is a 24ms hop, while L3 Chic to L3 Dal is only 19ms. Since they take nearly the same path and the speed of light is quite fast, my conclusion is that SL's Chic to Dal link is a bit congested. It seems to me that my speedtest issues are that L3 seems to be hot-handoff routing. This is the best conclusion I can come to with my limit Google-Foo abilities.
  18. Having an MTU above 1500 may actually result in a slow-down if path discovery isn't working correctly and your packets get fragmented. That being said, if everything is working, the MTU should auto-adjust.
  19. It seems there is some sort of artificial limitation, results are too consistent. I tired Chrome, FireFox, and IE, all up-to-date. Not any huge difference. Does Win7 need TCP settings tweaks? I get 114MB/s on my LAN via SMB file transfers, and was just getting 50Mb/s from a single download for several minutes strait on Steam(Civ5 Gold Upgrade was only $5).
  20. I have 50/50 rated Active Gig Ethernet with a ping to Level 3 Chicago that has a min of 10ms and an average of 11ms. I typically don't have issues downloading, but for some reason, I am getting something like 3-4Mb/s from these tests. This is what I'm used to. The spike going down was my HD thrashing from swap.
  • Create New...