In reply to a post by Daemon66:

TBB generally just like to claim that the catch-all 'congestion' is the cause without going into any details.

There are also differences in the way cable (DOCSIS) works compared to xDSL. With xDSL you efectively only have one channel for communications even though this is spread over a number of frequency bands whereas cable is more akin to MLPPP with data sent in a number of bonded channels. Cable needs to request permission to transmit on a channel which causes additional latency whereas xDSL doesn't. All of this adds up to very different characteristics for cable and xDSL for speed tests and is why ping/latency graphs also look very different.

Could you go into some more detail on this, please?

In reply to a post by Daemon66:

But to be honest you really don't need to worry about the single-thread speed unless it drops below what you actually need for a single stream, i.e. about 25Mbps for a UHD video (though most will actually need a lot less than that).

This too. I'm not sure people on 300Mb would be fine if they hit the download button at a website and it delivers about 1/10th the performance of their tier due to single-thread issues with VM.

It's not unreasonable for customers to expect better.

Edited by deleted (Thu 25-May-17 13:58:11)

More detail, that's what Google is for. I was just highlighting that cable is very different to xDSL so you should expect different behaviour.

I very much doubt anyone uses single-threaded in-browser downloads for anything very big these days so not many people are actually going to notice that it is slower.
Most browsers support multi-threaded HTTP/S and so long as one can run several concurrent UHD video streams or download game software through multi-threaded portals one probably isn't going to notice an issue. And I'm pretty sure that covers the majority of VM's customer's use.

Single thread performance generally tracks likely video streaming (note I said streaming NOT download) performance generally, and even then we see people with variable 20 to 50 Mbps single results who still suffer issues streaming on Virgin currently.

I guess its the different behaviour of cable, and thus should recommend xDSL to anyone who wants to stream video?

What? You really don't think cable is different to xDSL?

So far as I can tell, gaming is fine on cable as is streaming and VOIP. The only thing that the difference dramatically affects is single-threaded downloads. If only we had another cable service to compare with.

Are you bonkers? I don't think I ever claimed that DOCSIS was the cause of ALL VM problems, in fact I think I clearly mentioned that causes of issues are manifold.

In reply to a post by Daemon66:

More detail, that's what Google is for. I was just highlighting that cable is very different to xDSL so you should expect different behaviour.

If you could give a really quick outline of what the limiting factors are that'd be great. I know the basics but specifics on why it would be such a problem and cause so much difference would be really useful.

Usually when someone makes a claim the burden of proof lies on them, so it would be really good to know why you think the differences between cable and xDSL could account for all this 'out of the box' rather than it being something up with VM's implementation or hardware.

A UHD video stream is very unlikely to need even 25Mbps, typically they are around 10Mbps. If people who get single-threaded speeds in excess of 20Mbps are having issues it is not because of that. It is more likely there is congestion that is causing too long delays between data blocks, something your 'bufferbloat' metric could highlight?

Ah ok so people are not complaining about problems gaming, Skype not working, video constantly buffering and other issues such as modem being offline for a 15 to 30 minutes and once back speeds are slow?

So yes cable is a very different technology to xDSL and you might notice that for years no bad news stories on Virgin Media from me but only recently because things have become bad enough, i.e. outside the norms for what we see for cable users and that is after a month or two where things looked like they might actually be improving.

Just fell off my chair laughing, might need to lobby Netflix as they are clearly wasting bandwidth and putting people off buying the UHD option with their minimum speed requirements.

So what should I do, come on tell me what numbers I should tell Virgin Media customers, and a good reason for not telling them numbers already in the test?

Cable was designed for mass distribution from the headend out to the customers. The reverse path is a later add-on and relies on a scheduled request-to-send mechanism to stop all the customers trying to talk at the same time adding latency to acks etc. Each upstream channel does this independently. TCP throughput is very latency dependent. Multi-threaded throughput is increased if multiple acks can be bundled together reducing the average latency.

Likewise on the downstream each packet (not Ethernet packet) of data needs to be queued since it is actually sent to everyone connected to that same channel, the modem picks off the packets intended for it. These packets which can arrive on any one of the downstream channels in any order then need to be put back together into Ethernet frames before being passed onto the router (or the router part of the SuperHub). Again there is additional latency added because the parts may arrive out-of-order but only when all have arrived can the data be passed on. This is why latency is higher and subject to more jitter on cable.

Edited by Daemon66 (Thu 25-May-17 16:00:20)

What? Am I talking english here? Of course people have problems, that doesn't mean that when cable is working it can't handle streaming, gaming, VOIP, etc. [censored]?