Netgear CM2000 Modem, Asus RT-AX86U Router and the Comcast Gigabit offering
About a year ago, I saw that there was a new offering from ASUS (RT-AX86U) that supported multi-gig ethernet and the new AX WiFi standard which my iPhone SE’s (2020) supported and decided to upgrade both my router and cable modem simultaneously to support the 2.5G ethernet and AX WiFi standard with the hopes of improving my latency on my WiFi router with some bump in battery performance on the iPhones (negotiating the AX band improves the iPhones WiFi power draw) and simultaneously future proofing my networking for when Comcast supports gigabit in my area. After doing some research on 2.5G cable modem offerings I decided upon the Netgear CM2000 to satisfy my gigabit/multi-gig DOCSIS 3.1 cable needs to accompany the Asus RT-AX86U. If I was going to upgrade the router to 2.5G capable ethernet, why not the cable modem as well? Well, unless you don’t already have gigabit or multi-gigabit… Or do you truly need it to see benefits? From my reading there is anecdotal evidence to support that in fact newer cable modems, even with the same parts inside may have better performance only because the providers (Comcast and others) have newer firmware on their concentrators to support newer-released cable modems even if they contain the same chipsets.
Coming from an already DOCSIS 3.1 capable Netgear CM1000 (single-gigabit capable), I wasn’t expecting a huge bump in performance on the WAN/ISP throughput side of things as at the time I had sub-gigabit cable-tier internet plan (300mb). Most benefits you’ll receive from upgrading your cable modem will be from the support of DOCSIS 3.1 to begin with, which both ensures you’re NOT on the dreaded Intel PUMA chipset and also all DOCSIS 3.1 modems support the full 32x8 channel bonding of DOCSIS 3.0 standard ensuring you’re negotiating the full pipe in addition to supporting the ultra-wide OFDM channels of 3.1 (if available from your cable provider in your area). The CM1000 ticks the boxes already… Doing some reading on the available modem chipsets, the CM1000 and CM2000 are identical except for the ethernet adapter which appears to be a newer Realtek 2.5G NIC… Really without having a multi-gig plan, there is no sense to getting a CM2000, or is there? We’ll get to that in a minute.
After plugging in the CM2000 to my former Asus RT-AC86U and giving the MAC address to Comcast to provision it (took about 10 minutes), I didn’t see any improvement in benchmarks I was getting off my router directly (RT-AC86U) between it and the CM1000. My theory held (for the moment) that the Netgear CM2000 (and other 2.5G ethernet or dual 1G-NIC bond capable ethernet DOCSIS 3.1 cable modems ala CM1100/1200) was for future proof only if you already have DOCSIS 3.1 capable modem like the Netgear CM1000 or Arris Surfboard SB8200.
Swapping out the Asus RT-AC86U router a few days later with the newer Asus RT-AX86U when it arrived didn’t improve my internet throughput speed tests (download/upload remained the same, as expected). I did note that my browsing experiences on all devices were more snappy though. Web pages loaded faster and more smoothly on both my iPhone SEs and my MacBook Pros on the newer Asus RT-AX86U vs the older Asus RT-AC86U. A touch faster on my Roku TVs too, but not substantially like the MacBooks or iPhones. Doing some actual WiFi throughput and signal testing on directly on the iPhones (non-AX iPhone 7’s that I donated to the kids) and MacBooks on both 2.4ghz and 5ghz at differing locations I found the older Asus RT-AC86U was just a hair better in terms of signal strength and throughput on the non-AX bands than the Asus RT-AX86U, but the difference was nil compared to the improved latency on the RT-AX86U alone and obviously my iPhones SEs (2020) which support the AX standard see improved bandwidth and battery efficiency more than cancelling out the nearly immeasurable difference in signal/throughput on non-AX bands. All to say the Asus RT-AX86U is a true upgrade over the Asus RT-AC86U which is already a phenomenal router in terms of speed/distance, stability, features and price. That’s part of why I chose the RT-AX86U; you get AX standard support, while preserving most (but not all) of legacy bands performance which is a common tradeoff when upgrading to AX capable WiFi routers, a tradeoff you shouldn’t have to make… This is an important metric (non-AX performance) to pay attention to when choosing an AX WiFi capable router as many devices are still not up to AX standard as of the time of this article even though the AX WiFi standard has been around for some time. The newer processor on the RT-AX86U is paying dividends with the improved latency / responsiveness and the tried and true design of the Asus RT-AC86U/AX86U/AC68U still strikes a balance between 2.4ghz and 5ghz performance across supported WiFi standards making the RT-AX86U “the” router to get for balanced results if you want the best of the best so to speak, but aren’t a corner case somehow, example, having all AX capable clients would be such a case.
I should note, I stream my home videos from my Synology DS218+ (connected directly to the router over ethernet) to my Roku TVs, and the RT-AX86U did improve the streaming stability, which is notable for 4K content which my older Roku TVs struggle with and even my newer Roku TV can struggle with depending on peak throughput of the content when at times the video is delivering more mbps than the TV can keep up with but the Asus RT-AX86U appears to be helping negotiate this, presumably through faster fulfillment of not necessarily airwave throughput, but time to respond to next buffer request IE better latency locally. Also off-topic, but the new DSM 7.x OS for my Synology (which came out couple months after I upgraded to the Asus RT-AX86U), did substantially improve throughput for my SMB transfers between my Macbook and it, and, my 4K content between the Synology DS218+ and my newer Roku TV stopped stuttering afterwards. I do recommend upgrading to the new DSM if available even though some software isn’t compatible with the newer DSM, the benefits outweigh the loss in my book.
For the next few months this was it, improved latency locally for fulfillment of home movie streaming from my Synology DS218+ and Roku TVs and improved web loading responsiveness both a result of upgrading the Asus router from the RT-AC86U to the RT-AX86U and presumably due to the newer faster CPU in it with some benefits to my AX capable iPhones. The Netgear CM2000 wasn’t an improvement over my Netgear CM1000 as I didn’t have a multi-gigabit plan, as expected. Well, I should caveat I did detect a 2% average difference in throughput via repeated testing between the CM1000 and CM2000 over several days at random times but that 2% could just be sample variance. I won’t even waste my time posting the data; the internet felt the same between the CM1000 and CM2000 on the RT-AC86U. The RT-AX86U on the other hand, the internet felt different (faster) in loading.
This however changed when last month I received a phone call out of the blue from a “scam likely” per my iPhone, which for whatever reason something said to pick it up, so I did. I’m glad I did. It was a Comcast sales/retention rep that wanted to discuss my plan which was going to convert from a promotional rate to a standard rate next month. She tried to upsell a cable TV and internet bundle at a promotional price to replace my existing internet only plan at promotional rate. I declined and indicated that I would play the game and switch to AT&T for a year. In making small talk though I asked about upgrading to a faster internet plan, she indicated that she could give me the next fastest internet only tier at promotional rate which was cheaper than my existing internet plan at standard rate and faster service. After pausing and doing some research on the internet while she was discussing to compare apples to apples of AT&T service and price vs Comcast’s proposal, I just happened to ask if gigabit service was available in my area yet (I check monthly, and had just checked days ago on my internet portal which indicated only up to 600mb was available), she indicated it was available, and further offered to give me gigabit internet speeds for the 600mb promotional plan price and further added there was no 1yr commitment should I cancel. That sold me, if I wasn’t happy, I could switch anyways, without penalty. I agreed. She indicated the changes would be immediate and that there would be a restart on my cable modem, there was, and while I was on the phone with her I tested my internet speed, 960mb/sec, gigabit essentially. Just like that. Win-win, I got faster internet, at a cheaper price than switching to AT&T (as they make you rent the modem, whether you need to or not), at gigabit speeds…
Note the 360mb/sec down before and after the Gigatbit plan upgade, 960mb/sec range.
Afterwards, I noted as expected, the internet was only a tiny bit faster in real life use both on the Roku TVs, my laptops and iPhones; as I expected the gigabit plan was overkill for my needs and all the real benefits were in getting all my DOCSIS channels negotiated via having a DOCSIS 3.1 modem to begin with (as I’ve previously discussed in other articles) which mitigates cable traffic congestion problems. However, comma, I noted after repeated testing I was hitting 960mb/sec on the download (not 1000mb/sec) and 40-44mb/sec on the upload. From recollection of my fiber training, ethernet has about a 4% overhead, which amounts to around 960mb/sec for a gigabit connection. But, I also know that Comcast over-provisions links by 20%. If you pay for a 300mb plan, you can get up to 360mb. So why wasn’t I getting 1200mb/sec on the benchmarks? I gathered it was due to there in fact not being a 2.5G gigabit link between the modem and router. Checking the status tab confirmed that both the ports to my modem and Synology were being negotiated at 1G, even though the NAS was plugged into the 2.5G port (probably because the NAS doesn’t support 2.5G), which I assumed both the WAN port AND the 2.5G LAN were both 2.5G capable. Nope. Only one port on the Asus RT-AX86U is 2.5G, you get to choose what gets it. After digging on the internet, I found that you have to go to Advanced Settings, WAN, Dual WAN, and change the primary WAN dropdown to 2.5G WAN to enable the 2.5G port for WAN use (confusing, yes). I then connected the Cable modem to the 2.5G port and sure enough, the Cable modem changed from blinking white to blinking blue for multi-gig detection, and, the router now shows WAN port as 2.5Gbps on status. Testing the speed also revealed up to 1249mb/sec down afterwards breaking the 960mb/sec cap before. I noted despite the small change in going from 960mb/sec download to 1249mb/sec, which I wasn’t tapping anyways to it’s full potential, did however seem to improve my internet responsiveness. Upload remainder at 40-44mb/sec. Checking the ping and jitter with repeated testing showed those hadn’t moved the needle either. So what was going on? A theory… Some of my other previous technical reading back when DOCSIS 3.0 standard was new revolved around ways to optimize the latency vs throughput by predicative analysis of egress packets from the modem. Say what? Think of the truck (in this case outbound packets from your cable modem) waiting till it’s full before leaving the dock (going to the ISP concentrator), but, by waiting till it’s full, what if it’s a long time till more freight arrives to fill it? When do you call when to leave the dock for the day? How do you know if the freight in the truck already is time sensitive? This is what ISPs were looking at years ago and deploying various means to help balance problems with online gaming being latency sensitive vs trying to get the most throughput out of the pipe without creating additional congestion. I don’t know where those efforts went, but, what I was witnessing here, increasing the pipe to my cable modem from my router without there being a bottleneck and in fact I’d argue most website probably can’t deliver 1249mb/sec to me anyways, was improving the total fetch time of requests on my end. My theory: by getting requests to the truck so to speak more efficiently/faster, even though my pipe isn’t faster, the truck driver (logic on the cable modem) has more accurate data in which to base it’s predictive logic on resulting in better efficiency. All to say, in fact, the CM2000 might have resulted in faster internet, with the Asus RT-AX86U even without a gigabit plan had I set the 2.5G port up months ago when I got both. Alas, hindsight is 20-20 and I assumed both those ports were 2.5G and it wasn’t registering multi-gig only because I wasn’t provisioned for it at the time and it would’ve negotiated (2.5G between the modem and router) later when it was. Nope. You gotta move the connection to the modem on the 2.5G port and assign it to WAN use on the router to that 2.5G port. Lesson learned.
WAN Port Successfully on 2.5G
Breaking the 960mb/sec limit of 1G ethernet linkage.
After getting the 2.5G working between my modem and router, even though I’m still not remotely touching my 1250mb down, I’m noting my latency is more like the fiber to house offering I had (on a good day, as that was the problem with it, good days became rare towards the end of my time with Consolidated / SureWest making it inferior to cable in fact) when I lived in California, but, I was getting that on cable service. Pretty impressive to say the least. What’s my point? Couple points actually…
1.Bigger pipe to the cable modem and router appears to improve latency, even though my effective inbound/outbound throughput hasn’t moved the needle. By pushing data to the modem more faster aka the “truck/driver” over the 2.5G between the Asus RT-AX86U and Netgear CM2000, appears to impact the modem’s logic on when to “call it” to send a packet to the ISP improving response of web browsing. This could also impact online gaming, but, I suspect some of those algorithms on the modem may detect and optimize for this streaming-like activity so I can’t vouch one way or another, but it’s suspected that enabling 2.5G alone may improve results of latency sensitive contexts like VPNs and gaming in all probability.
2.You need to configure the Asus RT-AX86U explicitly to use the 2.5G connection and move the cable to it for it to be used or else you don’t get multi-gigabit. Not intuitive. This should be documented in bold text somewhere on the Asus support site. It’s not.
3.Having “just” a gigabit plan with Comcast, is multi-gigabit (1.25G) in reality due to default 20% over-provisioning; you need a CM2000 (or other multi-gig capable modem) and Asus RT-AX86U (or other multi-gig router / combination) to get the most of it. Just because 1G-2.5G plans aren’t available, only mere gigabit is available, doesn’t mean your CM1000 or SB8200 etc is cutting it. It’s not, at least for Comcast that is; you’re leaving 200mb/sec on the table. Now you do need to upgrade both your modem and router to support that 2.5G though for that last 200mb/sec, pricey. Upgrading a modem without the router to support 2.5G is like my NAS: my DS218+ can’t use 2.5G, even though it was connected to the 2.5G port on CAT6 cable at that.
4.The Asus RT-AX86U, although not any faster than my Asus RT-AC86U on non-AX bands, is significantly more responsive in terms of request fulfillment both locally and WAN/internet making it a worthy upgrade by itself, irregardless of internet speed used or constraints present.
5.Be aware of the options at the customer service retention agent’s disposal when renegotiating your internet plan. I’ve heard customer retention agents be it at an ISP level or other telecom providers (think cell service providers) have special options at their fingertips to enable things like gigabit plans at sub-gigabit promotional pricing. Your milage may vary. I wasn’t trying to achieve this, but I’m not complaining and I’ll be retained by Comcast for some time as a result as it’s a killer deal. The agent succeeded in keeping me in the Comcast family for another year.
Some other notables…
When I had Comcast install my service here at my new home almost a year ago, I personally supervised the technician install and watched the metrics on the DOCSIS link on her smartphone during the install; I insisted I link the modem directly without splitting at the PoP to maximize the link power; the technician proceeded to tell me that she knew her stuff and that too much signal is also a bad thing, and that actually it needed to be tamped down and put my link to my cable modem on the weaker 7db split to “fix” it. I decided to test this and before writing this article I moved my connection to the stronger 3.5db split from the 7db split to see if I could negotiate more links. Nope. In fact as the tech said, I deviated further from the recommended link strength, without improving my results (tested and checked). I ended up moving the link back to the 7db split. All to say I covered previously that too weak a signal via splitting the PoP (point of presence) can lead to loosing links you could have, but the converse is true, too strong a signal isn’t a good thing either.
On 7db split from the PoP
Note the Power Level being close to 1dB on the 7DB split; ideal.
Connected to the stronger 3.5db split
Power level now drifting at 5db “hot”
Despite the added signal power of being on the 3.5db split, only 5/8 uplinks negotiated. No change from the 7db split which also had 5 locked (not pictured).
Monday, March 7, 2022