tl;dr: The Intel Puma 6 and Puma 7 chipsets have a performance issue in a part of the packet processing that is specific to TCP/IP, which went undetected for awhile because "ping" packets are not TCP/IP. Modems that use the affected chipset will have high jitter under load (especially load consisting of many short connections), and will suffer denial of service when sent traffic with many different port numbers. Additionally, under a previous firmware (but according to a commenter on https://www.reddit.com/r/news/comments/7fjvvt/hundreds_of_mi... not on the most recent firmware), a timed process would briefly block throughput every 2 seconds, adding to jitter issues. DSLReports has created a tool which checks for this at http://www.dslreports.com/tools/puma6 .
Does anyone here know why these modems even care about TCP/IP or ICMP or any other protocol higher than the network and physical layers? I could go read the DOCSIS spec in detail, but thought these operated at a similar layer to Ethernet (OSI layers 1 and 2.)
Puma 5 shipped with just enough resources for 8 downstream channels. Puma 6 has same resources, but 16 downstream channels. Bottleneck is in hardware accelerating packet decoding/assembly, queue/buffer is too shallow and more packets in flight result in stalls. You can disable this hardware block and magically modem stops being susceptible to DOS, but max throughput goes down to ~60Mbit (opposed to theoretical ~800Mbit with ~2Mbit DOS potential)
There are people with Puma 6 with ISP using only 6 downstream channels and their modems work just great. 8 is barely acceptable (even puma 5 has problems on the test and can be DOSed by ~10Mbit), 16 is just plan broken.
DOCSIS "modems" are layer 3 devices as well actually.
The DOCSIS standard requires filtering by IP address and port [1] built in to the modem, instead of at the CMTS. This is back-asswards, as a hacked modem can ignore the filtering, but it's part of the spec.
There are multiple sources of jitter in the puma 6/5: more channels = more jitter, more flows = more jitter/complete failure at ~2000 separate active flows in a 1 second period.
I was under the impression that most rigorous network measurement tests used TCP (e.g. traceroute -T) because it was well-known that routers and other network devices along the path often had a separate fastpath for ICMP traffic.
> and will suffer denial of service when sent traffic with many different port numbers.
Yup, I've had this happen WAY to frequently with the DG1670 CableOne forced me to switch to (and I literally have no option for other equipment, they require this specific modem for anyone with a static IP). The modem will literally lose connection and reset itself when under load (which takes upwards of 10 minutes to boot), if I could put my old SBG6850 back in place I would do so in a heartbeat.
Both IPv4 and IPv6 have a spot in their headers to tell what type of traffic is being carried. This is a 1-byte integer called a protocol number.
The TCP protocol is protocol number 6. UDP is 17. ICMP is 1. All are carried over IP.
Both IPv4 and IPv6 have another spot in their headers to identify their version. This is where 4 and 6 go, but it's not the same as the protocol number field.
Similarly, Ethernet has a protocol identification field, "EtherType." IPv4 is number 2048 (0x0800); IPv6 is 34525 (0x86DD)
You are mixing different versioning systems. ICMP is protocol 1 inside Internet Protocol, where TCP is protocol 6 and UDP is 17 [0]. Internet version (as it named in rfc790) is defined with four first bits, numbers 1-3 unassigned and 4 is our IPv4[1]
Technically ICMP is layer 4 protocol, but defined as integral part of IP [2]
Judging by the picture on the main page, I thought it would be about them overheating and catching fire...
Until Intel released its Puma 6 chip Broadcom dominated the modem chip market. The Broadcom chip does not have latency or jitter issues.
As much as I hate them, Broadcom dominates the market because (as long as you don't ask for source code or Linux or datasheets...) their modems and other networking devices generally work. I'm guessing the cause of this jitter is because Intel is trying to do more in software vs. the dedicated hardware Broadcom uses.
The DoS on the "Security Issues" page is even more unusual, unless these modems are actually more than just modems and include a router/NAT too?
In what world do WLAN & Bluetooth generally just work? Those are pretty much the remaining bastions of horrible software support, and for no good reason whatsoever.
Even the graphics vendors have got their shit together. That's how hopeless Broadcom is.
bluetooth will generally "just work" when using two devices with the same manufacturer's chipset (this is not a rule by any measure though).
in my experience, many bluetooth hiccups are caused by manufacturers having alternate readings of the bluetooth spec, which is a massive, hard to swallow behemoth.
if you have two 1st party BTLE devices, you can expect to see bluetooth at its best. for example, apple airpods + iphone are probably as good as youll ever see a bluetooth pairing.
To be fair, AirPods + iPhone are using more than standard Bluetooth - Apple has their magic sauce on top to make things work super smoothly especially for the initial pairing.
You can still use AirPods with non-Apple Bluetooth devices. The pairing process isn't as braindead easy of course, but they do work. That implies they adhere to the standard at least to some degree.
I personally only use my AirPods with Apple products so I can't speak to whether they perform less well on regular devices, but, I somehow doubt it. Beyond pairing, it seems the only remaining secret sauce is in the relay between the two earbuds and coordination between the to for which one actually connects to the source device. None of that matters for the BT spec.
Outdoors, particularly walking down streets, may be a different matter.
Tricky bit about BT is that it is actually designed to bounce of surfaces to some degree. This because in its most basic form, that of a wireless earbud connection, it may not be able to penetrate the human body.
So to work around issues with devices being on different sides of a walking, talking, bag of water, it tried to make use of echos from nearby walls and other suitable surfaces.
End result is that in a environment that is constantly shifting, like out on an open street, performance will be hit or miss from moment to moment.
But when indoors and stationary i have been able to get a phone to talk to a tablet and earbuds, while the tablet also talks to a keyboard.
This is a bit of a rant, and a bit of an open question of "is there really nothing more I can do."
I'm more than a little furious that after being sold a product that explicitly fails under load, the best we have in terms of consumer protections is a class action that, if my history of class actions holds true, will accomplish little to nothing to either punish intel or compensate the consumers.
For the last year I've been putting up with network issues that I thought were a result of the house I moved into, when now I find out that the sb6190 I spent a silly amount of money on has been degrading an equally not-cheap internet plan for the entire time. And now my best approach is "deal with it, spend 200$ more."
Here's hoping consumers in the EU are more protected, for us in the US I don't have my fingers crossed.
The SB6190 is a $95 product right now. Plus, it's just DOCSIS 3.0. You can get decent DOCSIS 3.0 modems for ~$50 if you don't need maximum speeds. If you've got under ~300mbps connection, you could replace it pretty cheaply. For example, a Motorola MB7220 is about $50 without any discounts. Even a 16x4 is about $70. These are Broadcom based modems.
A note, not to distract from the validity of your points, but the sb6190 used to be 150 [0] (yes both prices are somewhat negligable in the long term but if I phrase it in terms of "I'm out a nice dinner" vs 2 I'm still pretty miffed)
That being said, thanks for the clarity on 3.0 vs 3.1, I had pulled up the recommended SB8200 since I had broadly heard positive things about Arris in the past and saw it to be ~190; didn't realize the cutoff was as high as 300mbps, so I should be fine with my 250mb down to get a cheaper one.
Not surprised to see a modem/router from Australia's biggest Telco on the list.
Australia's internet is in such a confusing state. At the moment a house could potentially get internet via one of ADSL1/2/2+, VDSL2+, HFC, Fiber, Wireless or satellite and on top of that most telcos are shipping woefully bad modems/router hardware with equally bad, gimped, and undersupported firmware to consumers.
Sounds like Norway basically. Thankfully the local powerco decided to roll out fiber in my area recently, and i think some 99% of the households jumped onboard.
Before then it was all about ADSL2+ (or some such) over a aging copper pair.
would this show up under an IPERF3 test? I would think so but just curious if someone with one of these modems can check. Also curious if this shows up testing between switch ports on the modem or if it's just IP outside the NAT.
The dslreports forums have been studying this for a year or so, and should have those details available.
From a practical perspective, if your modem is on the list, your apps and browsers will load pages slowly, with weird delays all over the place (due to DNS being Most Impacted by the Puma problems). Replace ASAP.
100% across the board with an SB6190 (HW ver 3, SW ver 9.1.93V, service from RCN). So… either the test missed an issue, or my SB6190 has been fixed. (Latest posts in the forum [1] seem to indicate that at least Cox is working with Intel on a fix.)
I’ll admit that I was initially confused by the link going to a picture of a burning modem, instead of an article. Stupid me. Just follow the navigation links on the left.
To me, it looks like a typical website design of the 90s: a left navigation frame, with the contents in the right frame. In fact, according to the <meta> tags, that site was made with FrontPage, as was common on that era.
If anyone doesn't know, always avoid Motorola/Arris Modems. I remember a few years ago Arris tried to update my modem to add ipv6 support. Ended up having 2 months of constant issues where incoming connections to ports forwarded would cause the router to drop all outgoing packets until the incoming connection closed... 2 months later Comcast rolled back the update.
Motorola/Arris modems can have 3 major chipsets (TI, Broadcom, Intel), each with a multitude of firmware and configuration updates delivered on a per-carrier basis, so this is a _very_ sweeping generalization. I'd recommend researching a modem much more in depth than just the manufacturer name on the front.
And this is a basic problem of the consumer networking market.
I know at least one box i have here are listed by 3 different variants on the OEM site, and the only distinguishing factor is a single letter on a single model code pinted on the sticker underneath it.
Checking the box or anything similar and you would be non the wiser.
I think it demonstrates sheer incompetence of the Arris takeover of Motorola and a lack of quality control with software updates. And applies to the company as a whole regardless of chipset.
