CA3LE

I'm trying to give you ammunition to go back to Spectrum with so I'm suggesting that you get ahead of them and check these boxes. If you're getting 250 Mbps on the Linear test, you're doing pretty good. Especially in a rural area. The one I was concerned about is where you're getting 40 Mbps. Something is preventing that install from performing. If you just booted that Windows 11 fresh install on that machine and now it's pulling 250 Mbps instead of 40 Mbps. There you go... something is holding that machine back, software wise. Don't expect linear to perform as well as multithread, it's very difficult to score like @xs1 does for example (1000 Mbps linear results). Look at mine, I get 400 maybe 500 if I'm lucky and then 800-950 multithread. What you're looking for is to establish that baseline on your best running machine, understand how your connection normally runs on TMN and if you see a huge variation (like 40 Mbps instead of 250 Mbps) then you know something is amiss. When you show me two hardwired computers on your network and one of them is running vastly different it often points to software. Or something else unique to that device like it's network card or ethernet cable. You very well may have deeper connection issues, I just want to cover bases. You've told me a lot about your situation but remember I wasn't there. I don't know all of the things you've done yet... and if you want to do have me be witness and give my recommendations then you need to understand that I need to see the tests performed a certain way before I can understand what might be happening. I know you've done some of the steps before but I wasn't there. If you can get all of your machines running over 200 Mbps in my linear tests I bet that your other connection issues will be gone. Everything will run better.

Well, I tried. It's been over 11 days when it says it's supposed to take 5. Paypal hates me for some reason. Oh well. I've been using Donorbox since this topic was posted in 2019. I like them or I wouldn't be using them. People have had reoccurring donations to TestMy.net with them for years on end and not a single complaint. They do exactly what they say they do. Like placing an order with any other trusted online merchant. My truly preferred donation method is Bitcoin (BTC) Here's a fresh address. bc1pl0vgsjwz5dg0zg35hat4tdjvuwyqpnswfqhf9vqxq7rfr09e2y5spwcsrx And actually you can do that with Paypal now. Buy some Bitcoin on Paypal and send it to the address above. Here's the instructions from Paypal themselves https://www.paypal.com/us/cshelp/article/how-do-i-transfer-my-crypto-help822 I recommend picking some up for yourself too. -- Make sure you take self custody! Not your keys, not your crypto. Holding BTC on a centralized entity defeats the original purpose, you don't truly own it until you take custody. Everything devalues vs Bitcoin over time. It's the most beautiful computer program ever written. The first truly limited supply asset in human history. Think about it. I knew about it back when it hit parity with the dollar (worth $1) but didn't start investing until 2015 when I took time to start to deeply understand it. I wish I had understood it sooner. I've been with it for every halving, bull and bear market since. Once you understand how it's programmed to do certain things and realize that it's the greatest computer system on Earth price before mass adoption is irrelevant. Dollars are irrelevant once everything is priced in BTC. If you donate to my BTC address it will never be converted into fiat. Ever. I'll keep it in my family forever. Dollars, not so much. See Gresham's Law "Bad money drives out good" -- people will continue to use their dollars. People will horde their Bitcoin. Take the Orange Pill - Bitcoin meme.mp4

The computer you set as "Work" was able to pull faster speed. TestMy.net Test ID : xSC9P4N5t Still leaves much to be desired but it's consistently 2-3X faster performing the linear test than the computer with the "Office" ID. TestMy.net Test ID : 7kCVzVEII So it doesn't appear to be a network-wide bottleneck. If it was you'd see very similar results. This is a really hard one. Let's focus on the PC that has the greatest drop, Office. I'd eliminate the operating system and all installed programs from the equation. To do this all you need is a USB flash drive. This is a little more advanced but you seem to be an advanced user. Download rufus from https://rufus.ie/en/ (you'll use this to create the bootable USB) Download Ubuntu from guide at https://ubuntu.com/tutorials/try-ubuntu-before-you-install rufus has an option to download the ISO for you, I personally prefer doing 2 steps. Open rufus, select device (the USB) and select the ISO image (rufus correctly fills in the rest) then click "Start" Now you have an Ubuntu live USB Reboot your computer with the USB plugged in. You may need to use the boot menu to boot from USB, the key you need to press on startup differs by manufacturer. Most common can be found at https://www.boot-disk.com/quest_bootmenu.htm -- some systems show you what you need to press on startup, others just show a splash screen. As it's booting select "Try" whenever it prompts you. (twice) Once the live USB boots up you'll have a normal desktop with Firefox browser with the option to install Ubuntu, completely isolated from your Windows install. Obviously you don't want to install it, only use the desktop for testing. Open Firefox, sign-in to TMN and repeat your testing. Keep in mind you're running on a USB so starting Firefox will be slower but once programs are in RAM they'll perform well for you. Note: I tried this myself just now and couldn't get video to properly display. Because the discrete graphics card needs proper drivers with a full install. To fix this power off and connect to the onboard video. After that I was able to load it up and test. Multithread: TestMy.net Test ID : teT8Ud7RN Linear: TestMy.net Test ID : UuMQH4oaw What I'd expect to see, exactly the same performance as my properly running Windows 10 install on that machine. First try, maxed it out. If you do this test and your speed shoots up all of a sudden... then we have something. After that you can use your Live USB to quickly repeat testing on other machines. There may be something software related, something in common between the computers that could be causing issues. It's common for TMN to pick up on this when other tests don't by the way. Using the method outlined above you eliminate those variables by using a known good configuration. Give it a try and let us know what you find. Note: Including the ISO download it took only about 10-15 minutes to follow my own guide and get test results.

The next test I would perform would be to remove the router from the equation. Connect your modem directly to one of your hardwired computers. You'll then need to power cycle the modem, unplug and wait 10 seconds and plug it back in. This connects the modem to the device mac address. When you return it to the router you'll need to repeat this. Repeat the test. Give the computer a different identifier so we can differentiate. Multithread off and on. I know you've already been through a lot but I think you're on the right track to getting it resolved now. Stick with me.

(from email thread) That's showing me that something is majorly amiss. If you can run the same set of experiments on another hardwired computer for comparison we can narrow down what's happening better. I recommend using Identifiers, if you click on My Settings a drop down with choices will come up. This will help us differentiate the results from each other. A comparison will help us determine if the problem is local to that machine or if it's network-wide. --------- I see the laptop tests you've been running since our email. Seems like you're getting the same there. I'll respond back on this in a bit. - D

Quickly enable and disable multithread, run some tests with it enabled and some disabled for comparison.

You're on the list. Hoping to have something to show you very soon.

After I get the beta released I plan on reworking the ranking. I think you can get what you're after by using the "full listing" search option. By default it shows a lot of noise. But you can modify the search to show only what you're after. It's a little tricky. First, I sort by upload. Then tell it to only show tables that have over 45K results by modifying the URL string in the browser. Specifically I'm telling it 45,100 results to pull up the more popular hosts. The tables in the database that tool is searching are pruned daily to a max of 45K results, for speed. So a host that has over 45,100 results shows that there's more recent activity on that host. Depending on when you query this, it may be different. The pruning and caching of the count are all done automatically on a schedule by the system. https://testmy.net/rank/hoststats.up/1/45100 Same search, by download https://testmy.net/rank/hoststats.down/1/45100 Once I shift my focus back to databasing I'll work to make that more simple and intuitive. One of these days I'll make all the data available and easier to digest.

Nice! Glad to hear that. 😎

I'll email you up.

It seems that they market themselves as an Internet provider, not an INTRAnet provider. I'd point that out to them. Pretty clear language on their website. Says "Internet" everywhere. May want to point them to the ftc.gov Truth in Advertising. I argue they should provide near the speed they advertise, TO THE INTERNET or they need to change how they market their product. They do say, "Actual speeds may vary." and then in your agreement (https://www.spectrum.com/policies/residential-internet-services-agreement) e. Bandwidth. i. Subscriber understands and agrees that Spectrum does not guarantee that any particular amount of bandwidth on the Spectrum network or that any speed or throughput of Subscriber's connection to the Spectrum network will be available to Subscriber. If you really aren't seeing the 1 Gbps speed anywhere and are seeing more like 500 Mbps maybe just downgrade to the package speed you're actually seeing to save money. In my area on Comcast it took well over a year or two after they released 'gigabit' before I ever saw anywhere near 1000 Mbps, anywhere. One thing to check and consider if you downgrade, what are the upload speeds with the lower packages. I pulled pricing from their website to the Green Bay area and couldn't find the upload speed stated anywhere. You have really decent upload, over 30 Mbps. I talked with an agent at spectrum I would need the exact address to be able to verify if the speed is available at that location. 300/10 upload 500/20 1000/35 Perfect, that's exactly what I needed. Thank you. You're welcome, was there anything else I could assist you with today? I think 20 Mbps upload is still very decent, obviously 35 is better. Again, if you really aren't seeing 1000 Mbps anywhere besides spectrum is $20 worth it to you only for the extra upload speed? For that price, I think it is. At the end of the day, if you switch to a 500 Mbps plan, the only difference you may see is in the upload. Then try the Gigabit plan again later and you may find better results. Personally, that 35 Mbps upload matters (though, my other options at the time only had 5 Mbps). I just stuck it out, over time Comcast got better and better in my area.

I'm running normally to Dallas myself. TestMy.net Test ID : xJEqAvBKE Switch your mirror to Florida or New York to compare. Might just be something with your routing to Dallas right now. If it's persistent I wouldn't ignore it. Variables are the same on my end, if you normally get higher speeds and they've dropped all of a sudden it's usually an indication of something. The fact that speedtest.net or other speed tests don't pick up on this doesn't dismiss anything. It's extremely common for other speed tests to report different than TMN. They don't calculate the results as transparently as I do and you really don't know what they're doing to arrive at their number. My new version, result shown above, gives you ways to adjust the result perception. Example, you can delay the start of the calculation or calculate based on percentile. These options are off by default but show you how the result can easy be altered without necessarily lying. Did the speed happen, yes... does it help you by only knowing the best parts of the result? I don't think so. So I always calculate based on everything that happened. Here's a test moments later with 50th percentile and graph delay enabled. Keep in mind, that's just 50th percentile, other tests could crank it up to 95th percentile or something for all we know. Below: Same test, same servers, same variables... except I'm ignoring a lot of important information. This isn't how I test but I put the option in there to help illustrate how perception can be altered easily. And if my connection were running poorly right now the difference would be even more pronounced. TestMy.net Test ID : Y8qkmq-yl If you'd like to get an invite to join the beta visit https://testmy.net/ipb/topic/34612-beta-testers-welcome/ First things I do when I get much slower than expected results... reboot the router, modem and computer. Then test again. - If it's persistent I'll try another browser (Firefox, Chrome, Brave). - If it's slower only in your preferred browser (Firefox) and not the secondary then I'd clear cache & cookies or Refresh Firefox Click the Firefox menu button , click Help and select More Troubleshooting Information. Then click "Refresh Firefox" Please let me know if any of this helps.

Should only take you 10-15 minutes. I'll make a video when I get to my console. Show you what I'm talking about. ... still reading your response. Just wanted to comment on that real fast.

I do, I just need to reinstate my account. Over a decade ago I received a really large donation, which then automatically drafted to my bank. Paypal didn't like that the money left without them having my SSN, but it was their system that did it! They locked my account requested my tax information, I've since provided the information 3 times. I can login, see everything... can even see my SSN inputted. I've sent them photos of my ID, SS Card and proof of address, for some reason they keep me locked out. It's as if they hate me. We'll see what happens this time. They have another new website, I submitted all the information they requested again. I have another PayPal account I can use if it fails again. Not that it really matters but I'd really like to get my original PayPal account back. You'd think they'd care more about one of their original members. I'll let you know the address after their review process.

And it looks like you're even worse when you don't multithread. TestMy.net Test ID : wfCW0g4qY You only have a few tests. Run some more tests with multithread disabled on the Colorado mirror and also test against New York and Dallas. Run a few tests against each, multithread off and on. I need more data to make a proper assessment. Please sign up for the beta, you seem to understand how TestMy.net currently works and I'd love to have you test my new systems. https://testmy.net/ipb/topic/34612-beta-testers-welcome You also may have a connection issue I'd particularly like to see tested with the beta. We can help get you up to speed. Please explain your network. Help us also understand how your computer is connecting to your network. Is it on WiFi or wired? What are the specs of the computer? Which modem and router are you using? Do you have another device you can test with? An Android or iPhone works great. Even better if it's another hardwired computer, preferably a fresh OS install or one with known reliable performance characteristics. The newer the better. Note: Separate these in the database with identifiers set in My Settings, then filter in My Results. Having these questions answered combined with the extra results from different mirrors will help us better understand what might be going on. Bottom line: You shouldn't have single thread running 8-10x slower than multithread. But we need more information to say that's happening for sure.

Since you voted (above) on this topic you'll get an invite when I'm ready. I know I've said this a few times already but I'm getting close. There are just a lot of details to a release like this one. Same as TestMy.net currently, works on all modern devices you only need a web browser. As expected, better specs return better results. I'm super excited to share it with you. I think it will help a lot of people better understand what we do differently here. There is no other speed test like it.

Wow, nice! $3912 is way better. $4463 per month saved. That blows away my record by a long shot. $54K per year or $161K over the 3 years you were initially looking at. $268K over the 5 years. Sweet. Anytime. I appreciate you using my services. I'm very happy to hear that. I look forward to you testing that new connection on my new test. Which can combine servers together on both upload and download to serve a connection like that. But I'll also look to get some 10 Gbps connected test servers online ahead of your upgrade so you can properly test the single thread performance over 1 Gbps. If you'd like to help the cause I accept donations. Just scroll down here in the forums. Your connection is already a beast, you'll really be tearing it up after that upgrade. Can't wait to see it and hear more, please keep us updated on the progress. Thank you. THAT IS why I do this.

I would ask them to justify the additional monthly charges. You said, So I was thinking that you had to come out of pocket for the upgrades also. And then be charged extra monthly on top of that. Then you said, Makes me think they're footing the bill for the upgrades. If that's the case, if it were me and I absolutely wanted the 2 Gbps I'd see how much of that work I could do myself. I'd also check the secondary market (eBay) for the hardware. What you need may be too new but you might research and find older models that do the exact same thing. Possibly the majority of the system can be built out this way, you may still need a specific component or two from them but you may be able to still save a lot. Pretty much just running a few cables and making a few connections. But part of what you're paying them for is piece of mind. They're going to be fully responsible if anything goes wrong and if equipment needs to be changed or upgraded for whatever reason they should be there for you. Is it worth that premium? I'm not you so it's hard for me to tell. If the ISP you're using has been reliable for you and you're thrilled with your speed already... I'd say it's not worth it. I'd also need to understand better how you use your connection. How much data do you transfer each month? If you transfer only 1-2 TB/month then you'll probably be totally happy keeping 1 Gbps. But if you're serving lots of data from this connection, running servers 2 Gbps upgrade might make sense. (size in MB) / seconds per month * 8 bits per byte = Mbps (1000*1000) / 2678400 * 8 = 2.99 Mbps So 1 TB/ month is really only a constant 3 Mbps. You're not putting hardly any load on the network capacity in that scenario. But if you're transferring 100 TB/month then that average of 300 Mbps is going to be much more noticeable on a 1000 Mbps connection, especially if your network has peak hours of operation. If I needed the extra 1000 Mbps, I'd only do that deal if the pricing was in line with the 1000 Mbps connection I'm already paying for. I feel like you should be getting a discount compared to your initial setup. Unless I'm missing something, the hardest parts of the setup are done. What do they have to do, swap out a few things?! And I bet the cables are in conduit, making it simple. Maybe the equipment lease is the bulk cost. You're basically being serviced and charged like a wireless provider would be. You've left the consumer realm, that comes at a premium. But to pay 3.67x more for 2x speed increase is steep. You said you pay $2275 now and the new price would be $8357 -- so 3.67x the current price you pay. Let's put it to scale. I have 1000 Mbps/40 Mbps and pay $100 right now. Let's say that my ISP (Comcast) offered 2000 Mbps/80 Mbps... but it was $367/month. Obviously the guy who created TestMy.net loves speeds but I'm not going to pay that premium, just not worth it. I could almost setup 4 of my current connections at the same price. So I'd tell them to stick that quote....... back in their pocket. If you're not using the connection to serve lots of data (let's say you use 5 TB / month) but just want the speed maybe you could work a better deal with them. Tell them, "Hey, I'm just a single household and just want a decent connection in my rural area, not a big business reselling the bandwidth. Can we work a better monthly rate with a data cap of 10 TB?" --- a number that gives you headroom to grow. I would ask, "Are the two connections load balanced or do they aggregate into a single connection? Can I make a single 2000 Mbps connection to a capable host on the internet like I can with my current 1000 Mbps connection or will I be limited to the speed of one of the two connections?" btw, interesting read on backhaul I found yesterday https://dgtlinfra.com/what-is-backhaul-wired-wireless-fiber-ethernet/ ...and that's priceless.

Wow, the lengths you've taken to get your connection. I imagine it provides some kind of return on your investment but maybe you're just an eccentric billionaire for all I know. To be perfectly honest, I'm not sure and can only speculate based on what you've shown. First, for the price, it had better end up as a single connection to your network. They're allocating 1x single core fiber across the setup. I'm curious how your connected between those points currently. Don't you already have fiber between those points? I'm not sure what BULLCS, WEBUCS and BUDECS are. Maybe you can shed light on that, searches turned up nothing. My guess is that the connection comes in from your tower, hits some equipment/hardware at the tower, then travels to more hardware at your building and then to your network. I think at the end of the day what you're going to see on your end is what appears to be a single connection. But you may only be able to transfer at the maximum of one of those lines if you're making a single connection to a host. In most cases, unless stated otherwise, they're still two separate connections and when you make a connection to a host the decision will be made which one to use. Some type of load balancing (Round Robin or Least Connection) will take place. If you're making multiple connections then they'll be distributed across the connections. Like I said, in most cases. But dude, you're an edge case. Probably the most extreme I've ever seen. You're not like "most cases". They mention a "Backhaul Component". This could refer to a backhaul within your own network OR they may be talking about a backhaul to the ISP. In that case your connection has more direct access to the spine (think of connections like nerves running through your body). The 2 connections may be backhauled and then combined on the ISP end. In other words, it transmits and receives as two connections but you'll see it as one. Again, this is me speculating. I don't know the tech they're using well enough to tell you for sure. I'd definitely get it detailed in writing from them before moving forward. Please update this thread with more information as you get it.

What speed do you pay for? The max you usually get is around 32 Mbps. That rating is comparing you against other users on AT&T, if you subscribe to a slower than average package you can safely ignore this. The new version doesn't have a star rating. If I add it, I'll make sure give an explanation of how it's calculated and improve the logic. Maybe have it weigh your own average in there or give you a way to input your subscribed speed.

Hi Aqueum, Yeah, maybe I should put the forum more prominently in the menu. Used to be front and center. The new version is web based. I never want to require you to install anything, ever. I don't think it's necessary. Especially now, with the advancement of browser technologies, opening native APIs to the web. Web applications can now access the GPU directly, I mean come on. All I know, I have no problem working within the web's framework. The capabilities today blow me away. With the right software and internet connection you actually won't need any native applications in the future. Been developing heavily on it, performs really well on every device I've tested on. I'm getting close. It's powerful and I think it's pretty amazing what it can do. All with just a simple, light web application. With over a dozen options you'll be able to control and tinker with the way the test operates, altering the flow and rendering. I made the options for myself in testing, I was going to stick with what worked best. But I feel it's better to just let you control it all. With default recommended settings you can revert quick and easy but still have control to experiment however you'd like. There are some pretty powerful option combinations that can really kick your browser into overdrive. Since you voted on this topic, you'll be invited by email when I'm ready for you to try it. For now, my release version (2018) still does the job. Many of the tricks the new version uses were discovered in developing the current version. If you have the bandwidth, you should be able to show it with TestMy.net. TMN has the bandwidth and dozens of servers that do only one job. You just might have to coax it out by altering how the test pulls the data, by using multithread. First, I recommend testing under your new username. Run a few combined tests, then click the "multithread" toggle at the top of the screen (e.g. says 'Multithread off' in my screenshot). Run a few more tests, you may see improved results using this method because it combines multiple downloads and multiple servers together (the new version does this for both upload and download tests). Then go to mirrors (quickly select at the top of the screen by clicking "Testing [locale name]" e.g. CA3LE Testing Colorado Springs, CO) Repeat testing on a couple of mirrors to compare how you perform to different locations. I pull only 400-500 Mbps using TestMy.net's Linear (single thread, transferring a single file) method, 700-900+ when I multithread. Other's like @xs1 regularly pull over 800-900 Mbps here using Linear. He used to only be able to do that with multithread... then one day, POOF! It magically bumped up. TestMy.net's servers and software were the same, pretty sure his computer and network were the same... then, poof, it just worked better. A variable had to change somewhere. TestMy.net Test ID : xwB2N5GI9.Onniz66P7 My opinion since the beginning, the single thread Linear result is the one you want score high on. Comparing the two gives you deeper insight.

Me too, hated that. Had AOL for a few months back in the dial-up days. Funny side story, related to TMN: A 15 year-old friend of mine was the master admin at a local dial-up ISP, he basically built and ran the whole thing for them and had been doing it for years before I met him. Gave me a lifetime free account, he moved on from there like a year later. But I ended up having dial-up until they stopped doing it like 10 years later. Helped me continue to real-world test dial-up here, for freeeee. The Sound of dial-up Internet.mp3 Came in handy on the road too. And they were always super fast! Always connecting at the max 56K! Ahhh, simpler times. Remember you'd have a 56K modem and AOL (USWest did this to me too.) would connect at 28.8K or 33.6K, "what the hell good is this modem if nobody let's me connect at 56K!?" -- 28.8K to 56K was a big deal, but AOL was big time overcrowded. Check out the 8 different types of dial-up sounds, some you may not have heard. https://en.wikipedia.org/wiki/Dial-up_Internet_access I think it was a 9600 baud modem that I used to connect to my first BBS. 9.6 Kbps or 1.2 kB/s. Now, the worst I ever see on my phone is 0.08 Mbps or 80 Kbps, 8X faster at it's slowest and that feels painful! Slower than that and it's not connecting reliably at all. At one time 80 Kbps was fast, it's all relative. One day we'll say the same about Gigabit. Kids will be like, "Great-Grandpa, what came before Exabyte internet? Mommy says you used to have Jibby-Byte internet!" "One point twenty-one JIGGA-WATTS!" 1.21 gigawatts？! - Dr. Brown & Marty - Back to the future.mp3

I don't mind giving it, just don't like typing it in plain txt... bots find it and spam. I sent you a PM, sorry I meant to send that after I hit submit on my last post. I removed your email address from your post because this is completely public. When they say 7.2 Gbps, they mean across the entire router. Multiple devices would need to connect to achieve that. 802.11ad Wi-Fi theoretically maxes out at 4.6 Gbps to a single device. So, if your provider was powerful enough you could maybe get up to 4.3 Gbps on wifi to a single computer. Your latency would be worse on wifi than on ethernet but you'd be cruisin'.

Now I want to make a little device that plugs into USB and when it's powered on plays dial-up modem sounds. So that when you power on your modem every day it sounds like you're connecting in the 90's. Cherry on top, "you've got mail!" Please send it, I'd love to read it. That router will have no problem doing 1 Gbps. It has 7x1 Gbps ethernet ports, hardwired to those connections will get your best performance. The nighthawk won't even notice it, won't break a sweat. If your connection is over 1 Gbps it can still be beneficial. You may be limited to 1 Gbps per device on your internal network but if multiple devices are pulling lots of data it can help. Most people don't have real-world use cases (yet) where they really need that. 10 Gbps has been available forever... but not even 2.5 Gbps LAN has really taken hold yet. It's still mostly 1 Gbps on the majority of all consumer equipment. The majority of my own home network is still 1 Gbps. I have a 10 Gbps network but it's really just for communication from my main computer to my trueNAS server. There was a while when we went from 10 to 100 then to 1000 Mbps, felt like every few years... seemed like we'd just keep iterating the standard but then it got stuck at 1000 and we're pretty much still there today, decades later. The enterprise (datacenter) world kept iterating. 100 Gbps+. My connection comes from Comcast's modem, into a 1st gen Netgear Orbi router... then runs 100 feet via existing CAT-5 (non-shielded) into a GS724T (26 port Netgear switch) then through a Microtik 10 Gbps switch to finally get to my main console. After all of that, I still pull over 800 Mbps on Comcast regularly. TestMy.net Test ID : iYtVDPhfC My router is at least a year older than the first Netgear x10 R9000. Your router isn't just greater than mine because of age... it's a straight up performance machine. So...... yeah dude, your router will work great. Feed it more speed. Make sure you vote on this topic to be part of the upcoming TestMy.net beta release. I'll be giving you some new ways to test your new found internet speed. By the way, I tested 10 Gbps on the 100 ft run. Wasn't able to get 10 Gbps but got a super solid 5 Gbps using the same equipment. So for anyone thinking they need to change wiring to get 10 Gbps... not necessarily true. If that run were a little shorter I bet it would do 10 Gbps. You don't always need special shielded cables. The CAT-5 in my home is from 2005, basic CAT-5. But even a 10,000 Gbps network connection isn't going to help you much if all of your computers and connection out to the internet have a 1 Gbps bottleneck.

Been a few minutes! Happy to see you. I think what they mean is that it's based on THEIR filter. The "filter" being 1000 Mbps compatibility, the SB8200 shows on Comcast as being juuuuust under @ 957 Mbps. I used an SB8200 for a long time, solid modem. I think it says 2 Gbps because it has 2x1 Gbps ports and has 2 downstream x 2 upstream DOCSIS channels. In theory you can pay for two connections with your cable provider but use only one modem. But it's still only suitable for up to 1 Gbps. I normally advocate for using your only own equipment and follow my own advice, but right now I have a Comcast modem. They only offer unlimited data if you use their modem. (maybe that's changed by now) I'm only using it for a connection. It then goes into a router which does everything. If you have to use their equipment, I'd only trust it to connect me to the internet. No routing, no wifi... personally, I don't want any ISP having any level of access beyond the gates. Hey ISP, I want you to give me a tunnel to the internet... nothing else. If I want more stuff, I'll get it myself. All we should ever ask is for one thing from them. An unaltered connection at the speed we pay for. "...but they can reset your password, power cycle your network... do it all on an app, make life easy. Come on man, trust them! You're paranoid!" --- having to hit the reset button manually or briefly unplug the modem is a small price to pay for privacy.