Superhub : My Network

[Kushan]

Originally Posted by MUD_Wizard

That's not it at all. The SNR targets/thresholds haven't shifted for 256 QAM over the years and 256 QAM has been used for a long time.

You need to define "A long time". I'm talking about ~5 years ago now and I don't know what the OP's timeframe is, nor if there are agents that still use the "old" guidelines (Wouldn't surprise me).

Originally Posted by MUD_Wizard

SNR thresholds are based on modulation, not whether it's Docsis 1/2/3 as such. Docsis 1 did introduce support for QPSK and 16 QAM modulations and Docsis 2 added 8-QAM, 32-QAM and 64-QAM.

256 QAM lower threshold ranges from 28 to 30 dB.
Add 3 dB headroom for local conditions and you get 33 dB.
VM will often quote one or the other of those figures, but they haven't moved.

I'm not disagreeing with anything you're saying, however all I can tell you is what VM staff were actually trained to look at. I'm not saying it's correct, just that's what the training told agents to look for. Modulation never came into it (it was very briefly glossed over).

Originally Posted by MUD_Wizard

Data corruption, visible in the form of post-rs errors, usually starts around 34 dB, but the noise errors need to come in clumps in order to trigger post-rs errors. If they're spread out enough then you'll just get lots of pre-rs errors and no or few post-rs errors.

However, to make things more complex different Superhubs display slightly different RxMER figures on the same circuit, which you have to take into account when giving advice.

Again, agents were taught to look at it slightly differently. They largely ignored post-rs errors for individual customers and instead only looked at the errors on a specific line/cable. They have (or had) a tool that gave a lovely colourful view on the errors in that area and a lovely arbitrary %. If the % was above a certain number, it was raised as an outage. If not, it was left to it (Ironically, the threshold was something like 30% but when a line was completely down, the tool only ever reported a static 25% - NFF!). For individual customers, it was T3 and T4 timeouts that mattered in this case (As well as power/SNR).

Originally Posted by MUD_Wizard

Btw, the next modulation down is 64 QAM, the lower threshold for which is 22 to 24 dB. So they won't be quoting that.

No indeed, they quoted 20dB in some areas as being acceptable (ex-NTL if I recall).

Originally Posted by MUD_Wizard

There are slight differences in power level ranges due to area, but the current range of -6 to +10 dBmV (downstream) generally holds, with some fine tuning by area. RxMER / SNR has an impact on how far outside that range you can go before experiencing problems.

CMTS manufacturers is a whole other conversation.

The power levels these days are definitely more consistent and it was aligned internally with the Docsis 3 rollout, however prior to that the range was much larger (I want to say -10 to +12 but I'm not 100%) and again, did differ by area on occasion.

If I had to guess, the guidelines being different for ex-TW and ex-NTL possibly stem from the merging of the two. I would not be surprised if both companies simply had different guidelines for their agents and Virgin simply decided to keep treating them separately until they got their ducks in a row.

[MUD_Wizard]

Originally Posted by Kushan

You need to define "A long time". I'm talking about ~5 years ago now and I don't know what the OP's timeframe is, nor if there are agents that still use the "old" guidelines (Wouldn't surprise me).

Virgin introduced Docsis 3 in 2009, with multiple bonded 256 QAM downstreams channels for 50Mb, however 256 QAM would have been available on Docsis 2 before that (but not at launch of Docsis 2).

http://about.virginmedia.com/press-r...test-broadband
http://forums.thinkbroadband.com/vir...fpart=all&vc=1

The number of channels bonded doesn't change the SNR thresholds set by the Docsis specifications. So, sometime after 2003 and well before 2009. Unless VM skipped Docsis 2 for some areas..

Of course most people will have been on Docsis 1 and gradually been moved over to Docsis 3.

Originally Posted by Kushan

Again, agents were taught to look at it slightly differently. They largely ignored post-rs errors for individual customers and instead only looked at the errors on a specific line/cable.

The flap list etc for area outages.

Originally Posted by Kushan

For individual customers, it was T3 and T4 timeouts that mattered in this case (As well as power/SNR).

T3 and T4 timeouts are upstream. It's SYNC errors and MDD's for downstream.

Originally Posted by Kushan

No indeed, they quoted 20dB in some areas as being acceptable (ex-NTL if I recall).

As I said, it's related to modulation. For more info see: http://community.virginmedia.com/t5/...4/td-p/2271297

Originally Posted by Kushan

The power levels these days are definitely more consistent and it was aligned internally with the Docsis 3 rollout, however prior to that the range was much larger (I want to say -10 to +12 but I'm not 100%) and again, did differ by area on occasion.

Power level ranges changed over the years due to aggregate/bonded downstream power and the number of bonded channels changing. See the primer linked above, it has a section devoted to it.

Originally Posted by Kushan

If I had to guess, the guidelines being different for ex-TW and ex-NTL possibly stem from the merging of the two. I would not be surprised if both companies simply had different guidelines for their agents and Virgin simply decided to keep treating them separately until they got their ducks in a row.

Sure there were differences as far as I understand it. Though not on the basics.

[Kushan]

Originally Posted by MUD_Wizard

Virgin introduced Docsis 3 in 2009, with multiple bonded 256 QAM downstreams channels for 50Mb, however 256 QAM would have been available on Docsis 2 before that (but not at launch of Docsis 2).

http://about.virginmedia.com/press-r...test-broadband
http://forums.thinkbroadband.com/vir...fpart=all&vc=1

The number of channels bonded doesn't change the SNR thresholds set by the Docsis specifications. So, sometime after 2003 and well before 2009. Unless VM skipped Docsis 2 for some areas..

Of course most people will have been on Docsis 1 and gradually been moved over to Docsis 3.

Again, you're not really listening to what I'm saying. Internally at Virgin, they have shifted the SNR and power level targets over the years. Again, it seems TW and NTL had different targets as well for whatever reason (Either different equipment in use, different standards, or whatever). The guidelines used to be a lot looser than they are now, so it may well have even been a cost-cutting measure, who knows.

Also, when Virgin first released 50Mbit, not all agents were trained to deal with it. There was a special "50Meg" team of agents and only those agents would handle 50meg calls (as a customer you had a special number, or if you phoned through to regular BB support, you'd get transferred into the "50meg queue"). Part of this training involved understanding different power levels and SNR targets (As well as the channel bonding). If the only difference was that channels were bonded, then the "training" wouldn't have taken a week and been done by all agents.

Originally Posted by MUD_Wizard

The flap list etc for area outages.



T3 and T4 timeouts are upstream. It's SYNC errors and MDD's for downstream.

And it doesn't matter. From the agent's perspective, they don't care if it's upstream or downstream, just that they're either going to send a tech to the customer or raise an outage for the area - so that's how faults are diagnosed, either it affects the customer or it affects the area. Beyond that, the responsibility ends for the agent so they were never trained to understand what was really going on or what those errors truly meant. That was the job of either the fault tech or the outage teams to determine.

Originally Posted by MUD_Wizard

As I said, it's related to modulation. For more info see: http://community.virginmedia.com/t5/...4/td-p/2271297

Once more, and I really must stress this, that's not how the agents were trained. The agents were not trained to check modulation and then assess power levels, they were simply told "If this customer is ex-TW, the levels are x and if they are ex-NTL, the levels are y". Again, I'm not saying this was correct (In fact I know later this was very very wrong as new equipment was installed in both areas) and I'm not disagreeing with your statements, I'm simply telling you that this is how the agents were trained to identify the "correct" power levels. There were 3 sets - TW, NTL and DOCSIS3. Also, not all agents were remotely kept up to date with this either, not all of them got retraining as the targets were adjusted.

Originally Posted by MUD_Wizard

Power level ranges changed over the years due to aggregate/bonded downstream power and the number of bonded channels changing. See the primer linked above, it has a section devoted to it.

Once more, I'm telling you how agents were trained to identify it, not what the actual reasoning is. And this is probably why different agents gave out different information over the years.

Originally Posted by MUD_Wizard

Sure there were differences as far as I understand it. Though not on the basics.

I was not present pre-merge so I cannot comment.