City Fibre

Apparently they were assured the range was correctly located and ready for use ahead of time and took that assurance as being correct. I suppose at least this time they asked, but I am a firm believer in ‘Trust, but verify’ and clearly no meaningful verification was done, despite claims to the contrary - you can’t argue with being geo-located in Iran, though TikTok not working is a bonus and no ebay would have a positive impact on my bank balance. Speed issues kicked in for me yesterday morning, we are up to about 100mbit today, on the basis that it’s been communicated, apparently CF requested it while migrating, and the timescale for it to be resolved is apparently just over a day away, it’s not worth getting bent out of shape over at this point.

I would love to think this could be the turning point, as they are making the right noises (moving things in-house, investing in the network etc.), but if I am honest, I did have a look at three other ISP’s yesterday, my wish list of multi-gig, decent peering, static IPv4 and ideally IPv6, DHCP, UK based CS and ideally a no equipment option seems destined to end with compromise. The UK obsession with PPPoE is just odd.

I noticed the reddit post, which was different to whats on their own forum, explaining that it will be their own equipment, so presumably they were leasing before, and of course that creates scaling issues, as previously they moved between different leased networks to grow. In theory this should be the last network rebuild that causes an outage if true, as moving forward they could scale it up internally.

I also think they have severe pressure on IPv4, there is a post where the yazi owner posted some subnet stats and they were very high utilised, so the activation of the new subnet may have been from that pressure, they dont use CGNAT at all now but also still dont have that many IP's (at least not on publically declared subnets).

I am inclined to agree with you that everything is done with good intentions, they are the one new small player who is actually actively investing and trying to improve their service, but they are fumbling along the way. Its going to come down to whether you can tolerate these sort of problems, and want to give them more time, or if its just too disruptive then move to someone else.

For reference I am getting congestion on AAISP over CityFibre now in the evenings, here is a graph from yesterday. It isnt like this every day, but its been a while since it has been flat all through the period, its usually either small spikes or something like this, will probably email them tomorrow, to see what they respond with, as this is now increasing in frequency. Speeds are still full line rate, so currently the effect is on latency and jitter, my monitoring shows jitter at points higher than Virgin Media.

 
Last edited:
CG NAT was ditched early this year around the time they announced the upgrading of 900/900 plans to 1.2/1 (then kept the 900/900 option). Reading between the lines, Liam got sick of dealing with complaints and tickets from people who didn't understand what CGNAT was, so everyone got an IPv4, and prices went up to compensate later on for new subscribers.

It’s been a rough 48hrs, yesterday evening speeds were down as low as 1Mbit, a few brief outages today, and the whole thing feels like a poorly communicated and at times chaotic plan. Simple things like validation of IP ranges not being done (again), limited notice of planned changes that resulted in an outage, the claimed 250Mbit turning into 1Mbit, much of that was avoidable or at least foreseeable ahead of time. The suggestion is CF required the speed limit to be put in place while migrating, and if it had worked as advertised at 250Mbit, I doubt any significant push back would have occurred for the few days involved, but it hasn’t, and the pitchforks and flaming torches are understandably being readied for the 18:00 deadline today.

If you look at BQM’s from other CF ISP’s, you often find commonality when they do ‘something’ or at peak times, sadly the communication relating to ‘something’ rarely makes it to end hsers.

Now if you’ll excuse me, I am off to sharpen my pitch fork and dip my torches ;)
 
I don't understand why migrating customers, presumably changing them from being handed over from the National network on an existing port to a new port connected to Yayzi's new network, requires such drastic traffic shaping. This bit has never been explained - is the CityFibre National network really low quality? Are they doing this to avoid going over a certain utilisation figure that would affect their costs with CityFibre, and CF are so poorly run that they can't be flexible at all on this when a customer is putting investment in? The line from Yayzi is that CF have asked them to limit the throughput which doesn't make sense, because it's no extra load on their network, and these speed limits are happening across the entire customer base including people that have already been migrated to the new network.
 
Just checked my Thinkbroadband graph and, for yesterday (10th), I had some congestion similar to what you're showing although not identical. Previous days the graphs were flat.

OOI is that a IPv6 address under the AAISP bit? I'm only checking IPv4.

I'm with LitFibre
Thats a hash I think for the graph, I do monitor my IPV6 as well, which looks similar to the IPV4 graph.

Hopefully AAISP get on to CF if its their side, they do have a rep for pushing these issues.
 
Last edited:
I don't understand why migrating customers, presumably changing them from being handed over from the National network on an existing port to a new port connected to Yayzi's new network, requires such drastic traffic shaping. This bit has never been explained - is the CityFibre National network really low quality? Are they doing this to avoid going over a certain utilisation figure that would affect their costs with CityFibre, and CF are so poorly run that they can't be flexible at all on this when a customer is putting investment in? The line from Yayzi is that CF have asked them to limit the throughput which doesn't make sense, because it's no extra load on their network, and these speed limits are happening across the entire customer base including people that have already been migrated to the new network.
Yeah I find that weird as an explanation. CityFibre being a company that really doesnt want to talk to end users, we having to go on what Yayzi says is happening, but it seems so odd, as whatever bandwidth is consumed on the new port is something that is no longer on the old port.
 
Back
Top Bottom