VM to gig BB in the UK
so where do VM to gig BB ethernet (over coax) in the uk and how much does this cost?
when are VM expected to release the patch?
Does this affect vm's other <300 mb users too?
Intel's Puma 6 chipset, used in gigabit broadband modems around the world, suffers from latency jitter so bad it ruins online gaming and other real-time connections. The semiconductor giant is preparing a firmware update to correct spikes of lag and bouts of packet loss that repeatedly flare up in home internet hubs in America …
If you're on a SuperHub 2, you'll be OK. If you get any package that has a Superhub 3, you'll be needing a patch which isn't expected for a short while. By gigabit broadband modem, we mean gigabit-capable; you could get a Homeworks 300Mbps package and get a Superhub 3 box and suffer from extra lag.
Superhub 2, black, loads of lights on the end and embossed VM logo on the sides.
Superhub 3: white, only one light on the end and has mesh sides for cooling.
I still have a SH1 and have only had that a couple of years, thanks to VM not upgrading my old Motorola Surfboard from 2001 until it finally died, even when requested.
diodesign, this Virgin media problem as the only real cable operator in the UK, is far larger than it would seem as they dont allow you to connect any non VM CPE kit...
and are Insisting that all vm customers that take the 'VIVID 200 Gamer' option or above as per the current 300Mbit/s package, Must take the one and only 'Superhub 3.0' Compal CH7465-LG (TG2492LG) & CGNV4 etc docsis3 box, as their only option, and apparently, VM tech/managers are refusing to lift the apparent VM swap-embargo on these SH3 and take the older Superhub 2 ACs as a fix today....
while its hard to actually pin down the exact chip spec and RAM (can you confirm and add the direct url) it seems that these intel docsis Soc are far to underpowered today, and dont even include generic intel AVX SIMD instructions in 2016, and yet the totally underpowered CPU core is actually doing all the UDP/TCP packet work etc alone....
is that a BOM problem today, as in all the world's Docsis3 cable operators are trying to save pennies , where a current and up to date (Arm Cortex?) quad/octacore with DVB/sat>IP ,UHD1/UHD2 rec 2020 ,HDR compliant integrated SOC's are required today ready for 2017/2020 Docsis 3.0/3.1.
I can't help but wonder if this related to the issue I have with my SH3 when in modem mode. It absolutely hates my RT1900AC (It was OK with my previous router (RIP)), I was getting random disconnects from t'interweb and everything pointed to the SH3. I even got the SH3 swapped because of it which made little to no difference. I ended up having to switch the SH3 back to normal router mode, disable WiFi and DHCP and stick my router in the SH3's DMZ to get it reliable. I might wait until the next FW update and give it another go.......
yes , this is also a known problem so called 'Hub 3.0 port flapping' as in even when you switch these VM SH3 into modem only mode, the same high pings,TCP,UDP etc exists AND you also get this other bug manifesting ...
The SH3's Port1 plugged into a ethernet port (ether1) on any router,where Everything works fine for about 30 minutes, and then the super hub shuts down its ethernet port for about 30-60 seconds, then brings it back up.
thats the other problem that,s still to be fixed, as relates to your SH3 being connected to some VM cards in their racks... so again you get disconnected from your game server when this happens, so making the existing 200Mbit/s gamer option rather lame and currently not fit for purpose if you are upgrading for the holidays this year it seems, unless all these related MS chipset fix's are expedited PDQ...
> Every article. Every.single. article. Whining about windows 10 in the comments regardless of the content of the article.
It's a side effect of Windows 10 installation... it causes the syst...er... user, to become stuck in an infinite reboot loop. Microsoft has been promising a patch er, since it was released... but I wouldn't get my hopes up. At least it's not a -4931337 error... also known as 'Smug Of Death' bug that plagues certain macintosh users. That one is just awful.
"Every article. Every.single. article. Whining about windows 10 in the comments regardless of the content of the article."
Why not? A big target is easily hit.
It's up there with DOS for simply wasting your time. I recall spending ages fiddling with himem.sys and emm386.sys to get games to work. Move forward a couple of decades and now I have to spend ages getting printers, scanners and other games working in new and amusing ways. Oh hilarious: after an update a driver is now disabled because "reasons". The start menu thingie keeps on reinstalling shit I've removed because "reasons".
Do I sound unhappy with Win10? Sorry. I use Linux myself. I seriously compare taming a Win10 install to maintaining a Gentoo install or perhaps Arch (Arch is precompiled.)
Yes I will whinge about it but only on other people's behalf
"now I have to spend ages getting printers, scanners and other games working in new and amusing ways"
Actually my experience with printers has been that they install quickly and automatically in a Linux environment. (...and Microsoft still hasn't figured out how to manage USB devices properly)
From the viewpoint of someone who maintains about 8 Arch Linux boxes at home [plus the wife's MacBook] and a few hundred Windows machines at work, I don't think there's any comparison between the two. My Arch boxes have been typical Linux, i.e. lots of mostly simple configuration work and a few head-scratchers followed by years of solid service with occasional WTF's (such as when CUPS dropped support for parallel printers.... grrrrr....) and the Windows boxes at work have been typical Windows, i.e. a nightmare to configure followed by years of misery and "random" problems.
[Expletive deleted] Whining about windows 10
I choose not to match your rudeness, but jeeze, did it ever occur to you that there is a REASON everybody hates it? I knew enough never to touch Windows at all after 7, and damn little after XP. I have been almost 100% Windows-free for over a decade, having jumped into Linux beginning 22 years ago, and do not miss any part of it.
Oh my god, the article is nothing to to with windows, or Microsoft it's a bug in some Intel code in router.
Do you all do this in normal life too?
Coffee shop : what do you want to drink?
You : I'll tell you what I don't want. Windows 10
Coffee shop: get out.
Seriously get some perspective, take up a hobby, walk the dog, anything but obsessing about windows, it's not healthy. if you don't like it don't use it, but please stop bringing it up on unrelated articles bringing the whole bloody site down.
The Intel I219V on new laptop was disconnecting every second. Only 10MBps mode worked. Switched off "autonegotiate" and tried Slave then Master. Only Master worked at 100Mbps or 1Gbps to my 3Com based 8 port switch.
So have Intel a problem writing drivers right now? Both Lenovo and Windows Update offered "newer" drivers with same problem.
The sad thing is Intel inherited the old DEC Tulip network chips which worked damned well. No clue what they did to the technically to reach this state.
I had two Fortville-based XL710 cards and two previous gen X520 cards. The Fortville ones (post massive recall) still had major firmware bugs that caused them to pull the firmware with NO way to downgrade (the upgraded made a backup but no way to restore it!), and both had issues with SR-IOV and their own embedddd MAC filters. Trashed them and went with Chelsio and never looked back.
> So have Intel a problem writing drivers right now?
Intel driver have always been crap. If you ever tried to do anything slightly out of the ordinary with the display setup it would shit the bed at semi regular intervals; Assuming it ever worked at all.
Caveat: My experience was with shitsets 810 thru 945 on Linux systems
Some Intel integrated display drivers are crap, I'd agree totally. They continue to exhibit a stupid Gamma bug in laptops so that light colours get washed out and disappear on external monitors connected to many laptops.
You open up the graphics properties window and move the Gamma slider slightly and the colours suddenly pop back to life again - go slightly too far again and the screen goes crap again.
However due to a fight between the Intel Graphics properties and Windows Graphic properties whenever the machine resumes from standby the settings get reset.
Intel acknowledge but say there will be no fix...
Intel has always had issues with drivers. I worked for intel break/fix back back in the early 2000's . We were getting loads of lap tops with errors for wi fi. issue will not connect at star bucks. Well sounds like a starbucks issue cause it works fine here at intel campus. Turns out it was not a starbucks issue but a drive issue. Do to a hiccup the Intel wi fi card will not connect to cisco AP on channel 3 ,6 and 9. The fix was and intel driver hidden away on the intel web site. Now if you used MS or the manufactures drivers the issue would come back. Now those was and Intel issue with intel solution for intel employees that very few people knew about.
Should have gone with an ARM SOC.
A bit difficult - cable is a duopoly - Broadcom (which uses mips for cable) or Intel. They both suck in their unique wonderful way - violation of specs, bugs obscure quirks, etc.
This is one of the reasons I have declined SH3 every time Virgin tries to shovel it down my throat. I stay with SH2 as it is the last CPE which can still do bridging. So I use the Broadcom chip within in the most stupid mode possible - as a bridge and hook it up to an OpenWRT. No intelligence allowed (as there ain't any - it's cable).
Could've been worse; at least it's fixable via software update. I remember severe growing pains back in the day. A lot of 28k-33k-56k dialup modems had far worse bugs that never got fixed. Cable was plagued by latency & congestion problems when it first became popular.
25ms jitter, though... ouch. I've seen ping times lower than that.
We dont know yet if it can be fixed with software. Im waiting on the firmware now. However at least one tester shows the issue is not fixed. Im the DSLReports guy Xymox1 who found the issue and the guy Chris mentions in the story. Those are my plots. So.. We dont know yet if it can be fixed in firmware. Its a stay tuned moment.
With probably 99% of all useful 'net content being good old 7-bit ASCII text, who can read faster than 19.2K/sec? Even today, I still connect over dial-up from my property in rural Mendocino county about 15% of the time ... sometimes at speeds as low as 2400 bps (fog and aging, cracked, dusty cable plant makes for bad signal/noise ratio). The low speed doesn't seem to affect my "internet experience" much ... And that's barely 200 miles by road from Silly Con Valley!
Mine's the one with the Telebit Trailblazer in one pocket & Kermit code in the other ...
Shame on Intel, where standards seem to be slipping year on year - but anyone buying anything from Virgin should expect no better.
You're talking like everyone had a choice. I wouldn't call opting for 2Mbit/s (at a push) ADSL2 a choice compared the VM's cable service (up to 300Mbit/s) a choice.
For what its worth, older SH in modem mode works pretty reliably averaging under 25ms latency to most MMOs (servers quite a long way from UK as well). Sure of course VM screws up its service occasionally and their customer service is as stellar as ever, but most of the time the cable modem service is very good.
Glad I dont' have SH3. It will probably take VM forever to release their patch once Intel has released their patch (if the issue is patchable via firmware that is).
Nobody does testing any more. Kids just don't want to do all that 'boring stuff'.
I remember when I started in IT the 'IT Guys' would spend months and months going over the code, testing it and fixing bugs. If it took a year that's what it took.
By the time I left that same corporation 20 years later it was a hard job keeping the 'kids' interested in any project from the beginning of the week to the end of that same week.
As soon as a new bit of work came in they all dropped what they were doing and wanted to work on that...for a week or so.
I think part of the problem is that modern companies don't like to use what we used to call 'man managers'. Those people that used to give out the work, check what everyone was doing and how they were getting on with it. Stuff just happens 'organically'.
Does it bollocks.
"I think part of the problem is that modern companies don't like to use what we used to call 'man managers'."
Your problem is clearly that your organisation's management are insufficiently familiar with (let alone competent in) the miracles of Agile, and DevOps, and (2018_hiptrend_goes_here).
Don't forget Cloud.
That seems to be the solution to everything  these days but quite how you will backup multi-TB databases to/from the Cloud when the Cloud provider starts getting uppity at more than 50GB of traffic/month/client.
 not forgetting DevOps, Agile and all those other here today, gone next week (After the mandatory conference or three) fads.
I agree with the lack of testing. Even writing the test plans and agreeing test coverage could take several months, then writing and proving the tests could take just as long.
One system I worked on took us FOUR days to run the tests. Lots of physical measurments of things like SNR, Db levels and crosstalk were needed in order to get approval to connect the computer system to the BT phone network (BT Voicebank). I suppose none of that is needed these days but back then it was kind of important not to mess up the telephone exhange operation.
"Your problem is clearly that your organisation's management are insufficiently familiar with (let alone competent in) the miracles of Agile, and DevOps, and (2018_hiptrend_goes_here)."
As already mentioned these are just another fad that some CIO read about in a book and pushes down.
Nothing makes up for someone being paid to crack the whip and make sure no one is slacking off and the work is getting done. Dress it up all you like in fancy names, folks need to realise they get paid to work, so man up, take responsibility and concentrate on it.
Don't worry about what the team in the next office are doing...
This implies that Intel did no testing,
I bet they did. But they just didn't care about jitter. It worked for most uses, the people who might be affected probably wouldn't have the knowledge, equipment and an alternative modem to to comparative testing, lets release to manufacturing!
Incidentally, for those who don't have an Intel modem, this problem of jitter is why playing real time gaming over wifi is patchy and unpredictable, albeit the causes are slightly different, and probably can't be fixed by a simple driver fix. If you've tried gaming over wifi, discovered it works most of the time and then goes pants, and retreated to a wired connection, then that's what Puma 6 users are enjoying over their wired connection.
well, it seems that at least virgin media over 12 months ago, took the many reported problems as found by the select testers, and simply binned it all, and just rolled out the SH3 on mass to bring us to today's massive SH3 problems as regards underpowered intel Soc cpu doing all the packet handling, instead of passing these intensive operations to the included packet engine, how good that engine actually is is a mystery and its yet to be proven as a benefit for the paying consumer, but its required ASAP in the UK as virgin media is the only viable option for faster internet OC.
You have a 25MB external hard drive?
I do actually… well, 20MB… external SCSI with a blown up power supply. Years ago though if I wanted to transport that much data, it was via a QIC-40 tape drive that plugged into the parallel port (and was supported out-of-the-box by the Backup program in Windows 95).
Yes, still have it, and it took many a journey to and from my high school to someone's home.
When I was at Stanford, early one Saturday morning a Grad student drove to Berkeley on his motorcycle & came back with tapes of the over-night build of 3BSD. Our Professor, visiting from DARPA for a couple weeks/months (a dude by the name of Cerf, you may have heard of him), wondered how the hell our VAX had the latest version of BSD already running (10AM-ish), when the Switched56 connected source code system hadn't completed the download of the source, much less started to compile it.
Biker's answer: "My motorcycle's latency might be sub-par, but it still has a much higher bandwidth capability than your network!". Cerf's reply? "Nice hack!" (Note that a variation of this quote is often attributed to Tanenbaum in 1996, but it was a fairly common meme around 1980.)
Just something to think about ;-)
And I've just received another letter from Comcast (old school, they screwed up my email when I moved places, haven't bothered to recover it) to replace my trusty SB5100 (and experience all the greatness Exfinity could deliver;). And I just couldn't care less. DOCSIS 2.0 is plenty fast for my overpriced Economy Internet. Plenty of headroom (even if I wanted to upgrade, I don't). Surprisingly I often see the lowest ping to game servers (and I do not host any).
There has been some OS-related mudslinging, but does anyone actually *know* (ideally with evidence) whether there is an OS (e.g. an RTOS) in the modem itself?
In the absence of a misbehaving OS (or maybe just badly designed bare-metail code, without an OS), it's hard to see what kind of management ingenuity could cause a failure mode like this one, *and* let it reach paying customers.
"There has been some OS-related mudslinging, but does anyone actually *know* (ideally with evidence) whether there is an OS (e.g. an RTOS) in the modem itself?
The VM superhub, at least, has a web server for management/config etc so I'd be surprised if there's not some sort of minimal OS in there.
'The VM superhub, at least, has a web server for management/config etc so I'd be surprised if there's not some sort of minimal OS in there.'
As the superhub is more than just a modem (My old Ambit modem also has a web interface, probably 'hobbled'), the worry then is, considering what the onboard OS has to manage, how secure are the superhubs? and are we just waiting for the announcement that some clever fscker somewhere has borked whatever security mechanisms are implemented on the buggers.
As to this latency issue, until they fix it this is as good an excuse as any not to currently accept their generous offer of a SH3 upgrade from the old Ambit 256 modem that's currently running my connection, and which has been doing so without any major issues for over eight years (with maybe about 5-6 hours total downtime during that period.)
sure, if you just spend some time looking....
in fact its an OLD linux, if you are to take the required ARRIS OSS releases as the normal operations see
as you can see ,
"The SURFboard SB6190 is a DOCSIS 3.0 modem is capable of download speeds up to 1.4 Gbps!
Last Updated: 2016-07-06"
however, if you look into the download you can see they seem to be saying there's also a TI Arm version of the SB6190 ,as we already know, the 32channel DOCSIS* 3.0 Intel® Puma™ 6 SoC are (underpowered and lacking even generic AVX in 2016,as the SoC cpu is already 3 years old today apparently) Intel® Atom™ Processor.
the question remains, does the latest Puma™ 6 SoC with later firmware than the above link
"OSS components of SB6190 Cable Modem
FW Version 9.1.93D"
also use an ancient 6 year linux build and Toolchain, without all the fixes and security advisories unpatched or even optimised for todays upcoming DVB>IP (as per the extension of the existing sat>IP kit on sale today) aka the SH3 and derivatives.
I've been complaining about these sort of problems on my connection to my cable provider (Ziggo in the Netherlands, unfortunately now part of Liberty Global) and they've been telling me I'm crazy, there is no problem and everything works as it should. Turns out their Horizon media box uses a Puma 6 chip... Unfortunately they're not exactly fast and efficient when it comes to rolling out patches either.
This post has been deleted by its author
About a decade ago I got a ringside seat at Intel's WiFi chipset development. They went about this development in roughly the same way as this one, buying IP (at ridiculous prices, IMHO) from older suppliers and then cobbling together their first silicon. The result wasn't at all inspiring. However, they have very deep pockets so by the time they hit the third or fourth revision things were working nicely. These days you don't notice these parts -- they just work.
Expect Puma to follow a similar path.
The spokesperson said "some users expressed latency concerns".
No, they pointed out a severe latency flaw. It was broken. A steaming pile. Provably bad.
Saying "concern" implies that the users were worried about something that hadn't happened, or that was unknowable.
Looking at a thread from September, the Hub 3 is referred by a single poster as (in reference to the Web interface) "slow", "super slow", "as if it's lost a fight with The Avengers™ and is trying to crawl home on its broken arms and legs over some broken glass", Wi-Fi as "poor. Very poor", and the Hub 3 itself as a "terrible pile of rubbish", a "piece of excrement", "an unmitigated PoS", a "dire PoS", "a PoS", an "awful PoS", "that PoS", and "this PoS kit".
Basically, Arris, the only "concerns" customers have is the vocabulary failure your equipment causes resulting in users running out of synonyms for "turd".
I certainly didn't call it a "concern" when I spoke to VM sales department.
I asked them if the serious latency flaws and port flapping issues were still a thing.
I was told there were no issues.
I agreed to the "upgrade".
I waited for the postman and then installed the new hardware.
I phoned Virgin Media. I complained. The packet latency issues were so bad my "ideal for home workers" connection was unable to cope with IPsec and SIP connections.
Eventually (within first 24 hours) I was through to someone technically literate with him running ping and mtr to my home server and me running ping and mtr-tiny to my virtual servers in Maidenhead and Manchester. We both saw the dropped packets and latency occurred at exactly same time each time. He said call back in 48 hours if the issue is still there and a second Hub would be sent out.
I was getting annoyed. I ran further tests. I tested the Hub's LAN IP. Saw latency spikes of >100 ms from one end of 2 metres of CAT6 to the other. Plugged cable into a Wi-Fi AP and ruled out the cable (latency spikes of 1.2 ms). Phoned up and got that replacement shipped out. In the meantime, the Web interface on the Hub ceased to exist and hard/factory resets failed to factory reset.
12 hours later I couldn't deal with it any longer. I phoned up and demanded to revert to my previous package. After getting through to the fifth person I spoke to I hit the 36 minute mark and my SIP call disconnected because of the port flapping issue. Phoned back up, port flapped while going through security. Phoned up again, and again. Was no longer on Homeworks 300+ but was on VIVID 200 with the Hub 3.0. Demanded to go back to Hub 2ac, put on hold, port flapped and call ended.
Called from my mobile (which until I "upgraded" had been part of my phone system), and I asked if I could have VIVID 200 Gamer with the 2ac. Was told no. Asked for clarification of what package I was now on having been disconnected from so many calls. VIVID 200 at £10/month more than I had previously been paying and billing couldn't tell me why. Eventually found out they hadn't removed the Homeworks add-on.
In 71.5 hours I had changed packages 4 or 5 times, changed modem/router 3 times, made a couple of dozen phone calls, had my billing date changed 3 times, had 2 pro-rata refunds and 2 month-in-advance charges, posted a long thread on Virgin's forum calling the Hub 3.0 "not fit for purpose" 3 times ("The Hub 3.0 Latency Spike and Packet Loss Issue"), and pinned a Tweet on Twitter (still pinned)...
PSA: Do not "upgrade" to @virginmedia Homeworks+ 300 Mb or VIVID 200 Gamer - the Hub 3.0 is a POS and not fit for purpose.
I certainly wouldn't call what I had "concerns". What I had was more closely "A PITA due to a POS I'd take a hammer to if Virgin didn't charge for non-returned equipment". Apparently Virgin value the Hub 3 at £100.
You can't call my latency graphs on my Virgin thread "concerns". One of them shows how bad things get as soon as you "upgrade" to the Hub 3.0, and another shows good things get as soon as you "downgrade" to the Hub 2ac.
Biting the hand that feeds IT © 1998–2020