Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[bogof]

I've just ordered 900/115 with Zen for a new FTTP install (CBT in neighbour's flowerbed at the moment, we'll see when it manages to make it to my front door).  Didn't really want to order FTTP through any BT group company for fear my current copper line might inadvertently get upgraded / removed.

Currently have 80/20 FTTC with EE.  Pings around the 7-9ms region from Norwich to 8.8.8.8, Unifi speed test and from the TBB monitor, and I basically get line rate up / down. 

I'm mostly after the improved upload, but hopefully will get the benefit of significantly higher downloads and not take a latency hit as I use VPN to my office. 

The only things that have got my heckles up slightly after a bit more reading is seeing reports of a few instances of people ending up routed via Manchester (which I guess will usually be a latency no-no) and download rates being quite fluid.

So what are folk's real life experiences so far with these ultrafast packages?  Am I right to be concerned over the routing with respect to latency?  There's a niggle in the back of my mind that it might have been better to go with someone London based and / or with perhaps more scale.

[Alex Atkin UK]

Unless you're a competitive gamer, I honestly don't see you noticing the difference between 9ms and 19ms, and its unlikely to be that much different even if you got routed via Manchester.

Not sure what you mean by scale, AFAIK most ISPs just feed to London so the latency gets worse as you go further north.  Its just the nature of a cost effective network configuration.

For me the killer feature with Zen is how they're big enough to have things like Netflix caches, while being too small to be forced to impose content filtering.  Though granted as I have Plusnet too I haven't noticed them do the thing I had issues with in the past where they incorrectly blocked whole domains instead of specific pages they had been instructed to, but its still a niggle in the back of my mind I'd rather not worry about happening again.  I don't think ISPs should be able to call themselves ISPs if they only provide "some" of the Internet, not all.

As for FTTP, I obviously can't comment though all reports so far are the latency should be identical to FTTC at the same ISP.  So the issue here is purely if Zen will have higher latency where you are or not.  Though again, I'd call it a none issue.

[re0]

I don't have FTTP, but I do have their Ultrafast G.fast package. So while I can't answer with an FTTP experience in particular, I would say your concerns are general to Zen's network rather than their Openreach FTTP packages.

If you get routed to Manchester, you can just ask them to change it to London. Happened to me, latency of 25ms compared to ~10ms with the previous provider (on VDSL2). I can't remember their reason for it to be routed this way, but after the change to London I was seeing the usual ~10ms. However, sometimes I find it can be as high as ~20ms even now. Just the luck of the draw when getting a new PPP session, but I don't notice any difference until I see the figures.

In regards to the "fluid" download speeds, I find that applies to the single-threaded downloads. For example, I might be able to get 300 Mbps throughput with multi-threaded downloads, but single-threaded downloads may be approx. half. Fortunately, most of my downloads are multi-threaded - and some of the mirrors I use do not always have enough bandwidth to saturate my connection anyway.

[j0hn]

As for FTTP, I obviously can't comment though all reports so far are the latency should be identical to FTTC at the same ISP. 

There's a little improvement between the 2, but nowhere near what I've seen suggested in the past.

From my experience (and a thread discussing this on TBB) there's about a 1-3ms drop between OpenReach VDSL2 and OpenReach FTTP, via the same ISP with the same routing.
It's roughly 2.1ms (it varies slightly) lower latency on my FTTP connection compared to the same ISP on FTTC.

Not exactly game changing stuff though.

It's more of a difference than I expected, but way less than the 5-8ms improvement I've seen mentioned by others.
Perhaps they had interleaved lines.

I've been involved in a number of FTTC to FTTP upgrades and most had very slight (roughly 2ms) improvements in the base latency.

Edit: added a little info

[Alex Atkin UK]

The thing is once it gets down to 2ms, that's pretty much margin of error territory when it comes to the Internet.

The strange thing is, at the moment my Plusnet line is 10ms and Zen is 23ms, that's a pretty huge difference and I'm routed straight to London it seems.  They used to both be around 16ms though Plusnet was quicker for a few hops but equalised once on the Internet.

[bogof]

Unless you're a competitive gamer, I honestly don't see you noticing the difference between 9ms and 19ms, and its unlikely to be that much different even if you got routed via Manchester.

I don't game, but from what I understood there are some VPN setups where the round trip latency can become a significant factor, and have the effect of limiting throughput, though I don't understand (and was hoping not to have to learn) at which scale these things become an issue - is it 10 vs 20ms or 10 vs 100ms?

Not sure what you mean by scale, AFAIK most ISPs just feed to London so the latency gets worse as you go further north.  Its just the nature of a cost effective network configuration.

I guess I'm thinking about the quality and quantity of their peering arrangements they might have once you hit their London / Manchester base, as well as fatness of their connectivity locally.

As for FTTP, I obviously can't comment though all reports so far are the latency should be identical to FTTC at the same ISP.  So the issue here is purely if Zen will have higher latency where you are or not.  Though again, I'd call it a none issue.

Even if I didn't need the latency, in the absence of physical line based matters resulting in different latency on the last hop to the house, I'd consider such a wide difference to be perhaps indicative of general network sanity or capacity.

The strange thing is, at the moment my Plusnet line is 10ms and Zen is 23ms, that's a pretty huge difference and I'm routed straight to London it seems.  They used to both be around 16ms though Plusnet was quicker for a few hops but equalised once on the Internet.

If that extra 13ms is in the Zen network, then that to me would seem like a pretty significant difference and I'd probably be chasing them to understand it.  At the moment on my FTTC via EE I can VPN to Amsterdam and still have typically sub-18ms ping times to things such as Google DNS.

[bogof]

I would say your concerns are general to Zen's network rather than their Openreach FTTP packages.

Yes, that is a fair assessment.

If you get routed to Manchester, you can just ask them to change it to London. Happened to me, latency of 25ms compared to ~10ms with the previous provider (on VDSL2). I can't remember their reason for it to be routed this way, but after the change to London I was seeing the usual ~10ms.

Good to hear they're flexible enough to do such things.

However, sometimes I find it can be as high as ~20ms even now. Just the luck of the draw when getting a new PPP session, but I don't notice any difference until I see the figures.

The last time I had "luck of the draw" PPP sessions was around the time of Pre-BT / early BT days Plusnet.  In the last 8 or so years I've done a dance around BT / Plusnet / EE and back again (and repeat) serially abusing their various cashback offers to minimize cost.  Always around 7-8ms mark, rarely noted anything untoward (there were a couple of bouts of peak congestion nastiness just recently - I guess peak COVID - but other than that, and a line fault which was sorted, it's been good as gold)

Nervousness about prematurely losing my copper pathway and number though sees me reluctant to stick with BT/EE/Plusnet.

In regards to the "fluid" download speeds, I find that applies to the single-threaded downloads. For example, I might be able to get 300 Mbps throughput with multi-threaded downloads, but single-threaded downloads may be approx. half. Fortunately, most of my downloads are multi-threaded - and some of the mirrors I use do not always have enough bandwidth to saturate my connection anyway.

I'm not so bothered about flat out speed per se - to an extent 900/115 is a bit of an indulgence in tech for tech sake (everyone wants a nice fat test result though, right?! ) more about the extent of the variation and whether it is common to all on 900/115 services, or something Zen in particular are finding.

Of course this could equally apply to any network provider, I'm just singling out Zen as I've got an order in at the mo and if I got to the point of thinking better of it I could always pull the parachute on the cooling off.
 
This is my current EE FTTC 80/20 home connection (I run with smart queues in the router, without it would be much hairier):

This is what my Essensys serviced office connection across the road from my house looks like (it's a big fat fibre link in, and I'm on a shared 80/80 tier within that with own static IP, though I don't think I'm actually sharing with anyone from what I can tell).  Again, smart queues enabled.  Probably will have to upgrade the USG3 router once home FTTP is in, as it isn't fast enough to VPN at the higher rates.


I guess you can see why I'd treat anything resulting in much worse than these kind of results with utmost suspicion, given there's clearly no geographic limitations here.

[bogof]

It's more of a difference than I expected, but way less than the 5-8ms improvement I've seen mentioned by others.
Perhaps they had interleaved lines.

I've been involved in a number of FTTC to FTTP upgrades and most had very slight (roughly 2ms) improvements in the base latency.

I'm not expecting any significant improvement, I'm very close to my cabinet with a copper line that doesn't have any current underlying issues and reliably hits the buffers at the profile limits with margin to spare.  So maybe all things being equal I'd just make a marginal gain from not incurring the small penalty from the cabinet processing.  But for reasons mentioned about fear of my line being upgraded instead of a new provision, I'm going with a new provider outside the BT group, so to an extent bets are off on the rest of it.  It will be a bit annoying if I don't end up in as good a position as I'm in now.  But can't see any good way to get confidence either way.