FEEDBACK - Rogers Rocket Wi-Fi Modem Firmware Trial

[RogersMargaret]

Hello Community,

We are currently offering our users an exclusive opportunity to participate in an upcoming trial of the new firmware for our Rocket Wi-Fi Modem (CGN3ACR, CGN3AMR and CGN3ACSMR) and Rocket Gigabit Wi-Fi Modem (CGN3552 and CODA-4582). For details of this program, please see this thread.

This thread will be used for feedback regarding the firmware.  We've invited @RogersSergio, @RogersSyd & @RogersBob from our Networking team to participate in this thread.  Your feedback is very valuable and will be used to enhance the firmware before it is released publicly.

Thank you for your continued feedback and support.

[NOTORIOUSVR]

---

@Alex4161 wrote:
@Notorioususvr

I think the issue is more the router than the upload bandwidth. Before Rogers, I had a Teksavvy connection with a DCM476 modem on a 25/2 and eventually 100/10 mbit plan. My router was the Netgear R7000 Nighthawk. With this combo, I never had any issues of unusable internet.

---

It is not the router at all..

1. because the issue is present even with the modem in bridge mode (so no router, just modem > PC)

2. because when I have sufficent UL speed the issue is less apparent

3. I use pfsense on a very powerful computer.

I have no issues back in the day either... only starting ~6 months ago when they compeltely messed up the network after my area getting the Gig upgrade... I was on 500/20 in the interm, speeds were rock solid no matter what time of day., not once did I have any slow downs.

[RogersDave]

---

@jalalinator wrote:

I also have these two entries in my log, the latest MDD timeout was also on the 16th for me, experiencing only around 500mbps down in my area on the gigabit network.. chromecast is very choppy on and off with periodical buffering on this firmware for me, dave is there a newer coda model and if so what should I ask for when trying to acquire it? also if I do get it will I need to msg you again the mac address? cheers

---

@jalalinator, the modem is the same but it has additional protection against interference in it. It's really a case by case basis but looking at your modem, I see that you could benefit from it but there might also be a signal issue. I suggest you first swap the modem and then if that doesn't work, I can assist to have a technician visit you to clear other potential problems.

Dave

---

@Alex4161 wrote:

My question would be is have you guys noticed any issues with open ports and running out of connections?  I am sure that conventional logic would be to check the Rogers network, etc. and this might be suspect, but sometimes the Router can't handle the open port and I/O load and just chokes.

---

@Alex4161, we haven't noticed that but I'll keep that in mind for testing. I'll run some benchmarking on my side.

---

NOTORIOUSVR wrote:

I have the same/similar issue..

The issue seems to be the UL which is so limited for some reason that any saturation of the UL bandwidth makes your internet unusable.

Not one person from Rogers has been able to address this or even mention it so far.

---

@NOTORIOUSVR, saturating your uplink bandwidth will cause higher latency and drops, that normal operation and expected. What can possibly be improved is queuing to reduce drops but it won't help latency.

At the end of the day, if the connection only has room to transmit say 30 Mbps and you are pushing this 30 Mbps while at the same time try to send a ping packet, this ping packet will enter the transmit queue and will get scheduled amongst the other packets. Unless there is some QoS to prioritize ICMP traffic, it will always occur, regardless of the modem or provider.

[NOTORIOUSVR]

---

@RogersDave wrote:

@NOTORIOUSVR, saturating your uplink bandwidth will cause higher latency and drops, that normal operation and expected. What can possibly be improved is queuing to reduce drops but it won't help latency.

At the end of the day, if the connection only has room to transmit say 30 Mbps and you are pushing this 30 Mbps while at the same time try to send a ping packet, this ping packet will enter the transmit queue and will get scheduled amongst the other packets. Unless there is some QoS to prioritize ICMP traffic, it will always occur, regardless of the modem or provider.

---

@RogersDave

While I appreciate your reply, unfortunately the issue here is NOT as you make it seem.

The issue here is that I"m on a GIG plan with UL speeds that fall to barely 2 mbit!  This problem is not my fault and you're asking me to do Rogers job by limiting my download speed to say 10-15mbit on my router so that I don't get packet loss and high ping rates.

This problem is what I"ve been attempting to get fixed for close to 4 months now.  Yet no matter if I post here in the forums, or call in no one seems to be able to do anything about it.  

This is the kind of support I'm getting for a service that I'm paying for where I can't use the internet AT ALL 80% of the time when I am home.  I hope you can understand my frustration.

Modem stats are good, lines have been checked, technicians have been out to my palce... yet my UL speeds are rarely even close to 20 mbit off-peak times (at which times PL is minimal, pings still suck though), and generally in the 2-4 mbit area during on-peak times where if I even do a speedtest my internet is unsable because no other connections can be made due to insufficient upload speeds vs the allowed download speeds.

I'm at a point where I don't know where to go to get this resolved.  Rogers have left me with unusable internet for so long and no one seems to want to do anything REAL to fix it.

So I ask you @RogersDave What should I do as the next step?

[Hybrid_Noodle]

Wanted to give feedback on the .20 firmware as my CODA (Non-White Dot) has now been online with it running for almost 8 days without a reboot.

I am running the CODA in bridged mode to an ASUS router in which I have not seen any IP addressing issues as seen by other 3rd party routers. I am still in a 3.0 area so speeds are a bit inconsistent but nothing drastic as reported by other users. Modem is about 1-2 feet from the router and I have not seen any interference issues. Previously I have seen MDD timeout's on my modem which have since stopped since upgrading to this software version (There should be no correlation between the modem version and the MDD timeouts but wanted to state none the less). Online gaming (XB1) has had no noticeable issues with latency or connectivity.

I am interested to see what the next software version comes up with.

[jalalinator]

hey dave thanks for the reply, when over to the store and swapped out the coda for the one with the black dot, I am currently on the .13 firmware on this device and currently sitting at about 400/7 for the speedtest, anyway you could expedite the firmware push this upload is pretty weak I understand if you are busy, also you stated something about the signal levels if you can let  me know what I should do next steps it would be awesome thanks

[HeWhoDies]

Picked up the "black dot" router yesterday, currently have .20 firmware on it. 

I see no difference in wireless connection.  If anything, I have a weaker wifi signal with this one in terms of range.

On wifi, with a gigabit killer wifi card, on 5ghz network,  channel 149,  the most i've gotten was 250-300 mbsp download with a 4ms ping.

Hardwired, we're getting anywhere between 800-900 mbps. 

[jalalinator]

seems pretty fair out of an aio device to be getting those rates, tbh I wish I was getting your throughput on the wired connection, I am bridged and Im barely able to break 600 :), consider yourself one of the lucky ones

[blues_clues]

---

@jalalinator wrote:

seems pretty fair out of an aio device to be getting those rates, tbh I wish I was getting your throughput on the wired connection, I am bridged and Im barely able to break 600 :), consider yourself one of the lucky ones

---

You shouldn't base your available bandwidth by running a single session test.  Run the speedtest on multiple computers or multiple browsers at the same time and you will be amazed at the actual bandwith you have access to.  Also, your processor's abilty to drive the NIC close to Gigabit speed is a factor.  Today's dual/qaud core systems will push your NIC close to Gigabit speeds.  I have gotten a max of 960Mbps from an i7 based server.  Remember, the speed tests you run are to the host network, which may of may not have resources to offer to utilize that speed or bandwidth.  If all your content stays within Roger's infrastructure, then you've got the best experience.  Once your request traffic leaves Rogers, then you are at the mercy of the bandwidth availabe to you based on your routing path.  You are also on a shared bandwidth infrastructure media.  There are other factors which I won't go into here.  The DOCSIS 3.1 rollout should provide us with the experience we are paying for.

[BS]

Hi everybody:

This is just a bit of a side question about the whole speed issue (I understand the latency, wifi performance, wired performance, and all the other things that people are talking about), but just what do people use a GB of speed for.

I still function on my 60, where I have been for who knows how long - at one point I had my two adult daughters who did a lot of streaming and sending receiving HD pictures, and my wife and I are low end users to say the least.

When we had all four of us, we did run into issues of approaching our original caps, but we had those increased, but I never found our performance lagging ever.  I have never experienced faster than 60, so I have nothing to compare to, so other than watching the speed test dial approach the GB or even approach my own speeds, just wondering, what the whole issue of speed is all about.  I do know that part of the issue is the shared bandwidth infrastructure and how DOCSIS functions related to speed and estimation of maximum bandwidth to all shared connections on the node (I have been reading and learning), but as a user, should I even care about speed?

Just wondering.

p.s. Mods feel free to move this somewhere else if that makes sense - wasn't sure where best to ask this question, but if there is any group of people who would know, it sure is the crowd on here. Just let me know where you put it if you move it (I think you do that anyway).

Bruce

[VivienM]

---

@BS wrote:

Hi everybody:

This is just a bit of a side question about the whole speed issue (I understand the latency, wifi performance, wired performance, and all the other things that people are talking about), but just what do people use a GB of speed for.

I still function on my 60, where I have been for who knows how long - at one point I had my two adult daughters who did a lot of streaming and sending receiving HD pictures, and my wife and I are low end users to say the least.

When we had all four of us, we did run into issues of approaching our original caps, but we had those increased, but I never found our performance lagging ever.  I have never experienced faster than 60, so I have nothing to compare to, so other than watching the speed test dial approach the GB or even approach my own speeds, just wondering, what the whole issue of speed is all about.  I do know that part of the issue is the shared bandwidth infrastructure and how DOCSIS functions related to speed and estimation of maximum bandwidth to all shared connections on the node (I have been reading and learning), but as a user, should I even care about speed?

Just wondering.

p.s. Mods feel free to move this somewhere else if that makes sense - wasn't sure where best to ask this question, but if there is any group of people who would know, it sure is the crowd on here. Just let me know where you put it if you move it (I think you do that anyway).

Bruce

---

Fundamentally, no, you shouldn't care that much about speed. I would argue that an unlimited 60 (if it is existed) or the unlimited 100 is more than enough for essentially all residential users.

Why do people care so much about more speed, then?

1) Rogers has aggressively pushed the higher-speed plans because they make more money on them. If I download 10GB, if I have a 250 megabit/sec plan compared to the 100, that means I take up 2.5X as much network capacity... for 40% of the time. The remaining 60% of the time, my download used no capacity. Offering higher speeds means engineering a network for more 'peaky' traffic patterns, but on a large scale, it doesn't really cost more. (Note: if people were actually downloading MORE if they signed up for the 250 plan than the 100, that's a different story and Rogers' costs would go up. But I don't think anyone will do more because they have 250 than 100...)

It is worth noting that their marketing department has not had great luck making persuasive arguments for higher-speed plans, in my view. The idea that you need the higher speed plans for more people/devices in the house is... questionable... once you're at the 100 unlimited level.

2) Upload speed. That's always been a weakness of DOCSIS systems. If you want upload speeds, you will need a faster plan.

3) A sense of getting what you are paying for. If you pay more money (or even if you got that sweet deal on the gigabit bundle) for a faster plan, you want to see those numbers in your speedtests and your daily life. Not for any practical reason, but to avoid feeling you are being highway robbed.

4) Very large downloads. If you are downloading games from Steam or the consoles's stores, then the extra speed will probably mean you can start playing an hour or two before the download would complete on the slower plan.

Honestly, I think the gigabit plans are too fast for something mass marketed... and when people can't get the promised speeds (especially over wireless), they end up being very angry at the provider.

[blues_clues]

---

@HeWhoDies wrote:

Picked up the "black dot" router yesterday, currently have .20 firmware on it. 

I see no difference in wireless connection.  If anything, I have a weaker wifi signal with this one in terms of range.

On wifi, with a gigabit killer wifi card, on 5ghz network,  channel 149,  the most i've gotten was 250-300 mbsp download with a 4ms ping.

Hardwired, we're getting anywhere between 800-900 mbps. 

---

Try orienting your modem in a different way.  If it was upright, lay it on it's side or vice versa.  Based on the orientation of the internal antennae, you may get different results.  Mine was a gain of 4db (from -38 to -34) just by laying the modem on it's side, with the lights horizontal.  Go figure....

[BS]

Thank you @VivienM  That clears it up for me and confirms what I thought was the case anyway.

For my case (2 semi retired adults, one with a home business, and me, on disability), we don't watch very much netflix, and are basically pushing text back and forth in messages most often, and the occassional upload or download of pictures.

Very clear easily understood answer.  Much appreciated.  And I agree, if you are paying for that bandwidth, and not getting, it, I would be annoyed too.  I get annoyed when my 60 drops down below 20 (that is where I will typically notice it based upon my use), and a phone call goes in pretty quick.

Much thanks, Bruce

[jalalinator]

my desktop has a i76700K, even before on my 4770k I was hitting 930 down and that was back january-march of 2016 I was one of the first on the gigabit package paying 170+ a month for it and I always conduct the same three speedtest so my results are from the same servers each time including the rogers speedtest, my computer is more then capable of handling the throughput as is my network, you will never be able to reach 1 gigabit speed on a 1 gigabit card, just theoretically not possible. I am not doubting the rollout of 3.1 will assist in this at all, just the road here has been quite rocky.

[Double_K]

@BS I would add that the other purpose for the Gigabit service may be a prelude for IPTV.

When Bell launched it's IPTV product, the running joke was that you could only watch it if no one was using the Bell Internet connection at the same time.  So lesson learned was to ensure that the infrastructure can handle sufficient bandwidth for internet usage (as you say, 60Mbps is fine) + additional bandwidth for IPTV.  It's all about streaming Video now - to multiple devices in the home.

Rogers gets customers to sign up for its $99 Gigabit+TV plan to see what the usage is, and determine available bandwidth for IPTV.  Some people would call those customers "field testers" or "data points".  Behind the scenes, it enables Rogers to perform widespread analysis of how much video customers will stream, and a "production" view of load on the network.  It would be extremely costly to emulate 10's of thousands of customers load in a "lab" environment for testing, but if you can get the same data analysis in a live production environment using real customers - well you just can't put a price on the value of that data. (hence the big discount on the package price)

[BS]

Thanks @Double_K  That all begins to make more sense to me - hadn't made the link between IPTV bandwidth, streaming and Internet bandwidth and speeds.  Takes a while to get out of the old mindset of tv and internet being two different things (Roger's case - IP for Internet and currently QAM for tv modulation and tuning.

Maybe this explanation helps to understand why Bell is moving to not requiring a separate Internet package with their IPTV service - separating the two some how may make it all more manageable. And yes, I remember when there was the restriction on Bell that Internet and TV watching was limited by number of concurrent connections.

Never been an issue for me - I am one of those very old school people - One HD tv in the house, just bought it a couple of years ago, and I think we use about 40 GB bandwidth a month at most.

Hope people don't mind me hijacking this thread a bit. I appreciate the clarifications on what this discussion about speed all means to various types of users. Have also enjoyed watching the whole beta testing discussion and glad I don't do that anymore.

But I also beginning to understand better that just because I don't use much speed or bandwidth for video streaming, we are on a shared infrastructure and the capacity for reliable speed of large bandwidth to the whole node is critical for performance.  I do remember the days of speed deteriating in my neighboughood when we first moved into the area, and we only had Rogers for high speed.  It could be brutal during peak times when the kids and then parents came home in the afternoon. 

Thanks to all of you for clarifying.  Now back to my basic internet and TV watching. and back to testing for you guys. Technology has come a long way, but it takes a lot of testing and work to get it right.

Bruce

[HeWhoDies]

@jalalinator, Yeah, the wired connection is awesome, no complaints there, but unfortunately, we're paying for a downgrade on wifi. When I was on the 250 connection, with the CGN3, No problem getting 320 mbps down on our devices. Now we're lucky if we even getting 250. We've been considering bridging the router, or just downgrading to get the CGN3 back.

[HeWhoDies]

@blues_clues

When we run the speed test, we are maybe at 10 feet away from the router at most.
I tried a few different positions, and there is no difference in wifi speeds. Tried playing around with the different settings under the 5ghz tab and can't seem to get more than 250 - 300.

[VivienM]

---

@HeWhoDies wrote:
@blues_clues

When we run the speed test, we are maybe at 10 feet away from the router at most.
I tried a few different positions, and there is no difference in wifi speeds. Tried playing around with the different settings under the 5ghz tab and can't seem to get more than 250 - 300.

---

What wireless card does your test device have?

250-300 over wifi doesn't seem that bad to me unless you have serious 802.11ac gear... 

[blues_clues]

---

@VivienM wrote:

---

@HeWhoDies wrote:
@blues_clues

When we run the speed test, we are maybe at 10 feet away from the router at most.
I tried a few different positions, and there is no difference in wifi speeds. Tried playing around with the different settings under the 5ghz tab and can't seem to get more than 250 - 300.

---

What wireless card does your test device have?

250-300 over wifi doesn't seem that bad to me unless you have serious 802.11ac gear... 

---

I use an Intel AC7260 in a Lenovo T440s.  Here are some suggestions for everyone who thinks their experience on the Roger's infrastructure is sub par (not getting what they are paying for)

1. Run local tests on your hardware before you utilize external speed test servers.  I use iperf3 locally to determine the capability of my local hardware. iperf3 is very easy to use.  "ipef3 -s" on the server side and "iperf3 -c hostname" on the client side.  There are builds for Windows and Linux.  It will do TCP and UDP,.  Everyone should benchmark their local hardware first.  There are public iperf servers that you can also test against.

2.  If you want real world measurements, use testmy.net.  Enable multithreading on the tests.

3. Those speed test server results will always fluctuate.  Your results will vary based on the time and load of the target host.

Maybe RogersDave can allocate a test server for the beta firmware people? 🙂

[Double_K]

@RogersDave  Thanks for upgrading me to .20 early this morning.

Here are my results;

Plan: 100u

Model: CODA-4582 1A - White "launch" version

Mode: Bridge

Router: pfSense with no traffic shaping

Note:  I also ran the XBOX Multiplayer network test.  Anecdotally, it appears the download speed reduction & increased latency that I saw on .19 have been corrected (ie. .20 seems to have returned to .13 levels of 130+ Down, 10+Up, 72ms latency, 0% packet loss)

[Aim_eYacht]

Three disconnects and full reboots of the modem in the past 30 minutes.