Aquiss

[Rainmaker]

@Rainmaker glad to see it all worked out, enjoy!
now i just have to nod and pretend i know what you guys are talking about outside of the numbers.

might have to pick your brains with regards to my own setup

Yeah no probs mate.

After some fiddling and whatnot, I'm more than happy with the line. Everything's tight in the 10ms to 11ms range even under load, and I've lost hardly any headline speed. Most speed tests clock in around 850 to 860 Mbps and I'm more than happy with that considering. Without SQM it's a solid ~942Mbps but latency bounces up a little bit (still only to what VM consider unloaded latency lol).

It's absolutely night and day over Virmin; web pages load instantly with a snap, my servers are much more responsive (instant) and it's just so much nicer to use. Rock solid speeds and latency, day and night. Aquiss are the nicest/best ISP I've ever been with customer service wise, too. Well worth the wait!

 

[Rainmaker]

Jeez, this should give you all a laugh... I forgot I had smokeping running on the NAS. This is 10 days, before and after the switch from VM to Aquiss, and it speaks for itself.

 

[Rainmaker]

Here is my ye ole VM graph, please feel free to laugh as hard as you like

I'm assuming a ~200Mbps package? No bloat there, latency is nicely controlled. As good as VM gets. The initial spike is typical of VM, and appears to be an initial over-provision to make speed tests look better (it raises average speed...). Nothing you can do about that, but otherwise the graph is spot on.

 

[Rainmaker]

Yeah 250/25 I had to resign myself to another 18 month contract just before crimbo. I'm taking about 10% percent for QOS.

Edit: Added Graph
Here is the uncontrolled mess!
{img snipped}

Jeez, that 'before' graph is horrific. God bless cake and the people who made it, is all I can say! I don't want to derail the Aquiss thread too far into bufferbloat territory (not a dig, I started it!) but man, that's something else.

Edit: Try Dave Taht's server, it's a bit faster and more reliable (at least last time I tested the 'official' server). The address is london.starlink.taht.net.

 

[Rainmaker]

We should actually have a dedicated thread for Bufferbloat etc talk. It's a pretty interesting subject. Surprised with how bad pre full fbre is in UK that we don't already have one. My line is +500ms on upload if I don't have SQM. I'm getting tired of PPPoE making hardware selection and offloading that bit harder, particularly for SQM. "Oh you want 150 Mbps CAKE? Sure this £30 OpenWRT router will do it... but if you have PPPoE expect much less". Just annoying.

I'm in quite an interesting point in choosing a provider. Have the option (for awhile now) to go from FTTC to FTTP and BT are offering me a recontract for 500 Mbps at £29.99 which seems great but tbh the speed isn't much of a thing for me. 150 Mbps would be fine so considering Aquiss. I'm trying to stay off for as long as possible for YouFibre but it's hard (150 Mbps for £19.99 is perfect). No guarantee they actually service the street so may aswell switch.

From what I read up Aquiss have entanet as the backhaul and posts like this worry me: https://forums.thinkbroadband.com/otherisp/4701048-idnet-to-aquiss.html

"The word Entanet should be enough to scare anybody of. Twice bitten and never again!"

The traffic management is... interesting: https://ebilling.aquiss.net/knowledgebase.php?id=46-broadband-traffic-management

Basically reads like "pay us business rates if you want decent ping".

That has to be out of date. Martin said there is no prioritisation, and they operate neutrally. He also said I can expect the same rates 24/7 with no nonsense, and that's exactly what I've experienced so far. They've been going a long while now, back to ADSL, and I imagine (but stand to be corrected) that the page is a remnant from those types of services. There's certainly no priority to business service, I specifically asked this because if so I'd have bought a business package (as I have in the past). I was assured that nope, all traffic is treated the same - and he backs that up in your first link. Aquiss is only a 12 month contract, and their average customer retention is 7 years, so that says something.

e: Even in the posted policy (which again, I strongly suspect is outdated), it isn't bad. There's no mention of throttling BT or NNTP et al., just priority for VOIP and video calls. That's SQM 101, and I implement that on my own network. Those types of services are extremely latency sensitive, yet use minuscule bandwidth. Priority to VOIP isn't going to impact anyone's connection in any noticeable way, but it *would* severely impact the experience of someone making a call were packets delayed or dropped. Meh.

Edit to address:

Not to start anything here but it's the little things that annoy me with Aquiss. Just now I look at the main site and the footer still has 2023. That means they didn't bother to look or maintain the main site in any reasonable time. Also call me paranoid but anytime people talk about support tickets it seems to be Martin answering them. Why is the managing director answering support queries? It feels like a 1 man operation. If something happened to him I have a sneaking suspicion that your might not have internet for long. Their accounting don't seem to suggest much support staff based on numbers.

That's your prerogative, but honestly why *shouldn't* the MD be replying to tickets? It's one of the attractions for me - it's a small family business with a few employees, and the MD isn't hands off he's in the thick of it actually running the business and getting involved. When I had a small issue last week, I didn't press option 1, wait an hour and then speak to someone who asked me to turn it on and off again. I spoke to Martin, who knew exactly what was the issue, and who had it sorted there and then because he's been doing this for years and doesn't mess about. I fail to see how that's a negative tbh. As you said, you're not trying to start anything and you're entitled to your opinion etc, so no bother from me. But that said, as a customer I can say I know where my bread's buttered and I can't knock it at all - I certainly don't care that the date on the website is a few months out, I care that my broadband's decent and that the support is good!

 

[Rainmaker]

It was Martin I was talking too as well.

"We always apply the same offers, if running, for both new and current customers equally. So yes, any customers who have been with us for 3 years or more have been getting the 6 months too if they have been changing packages."

But when I asked for more information on that I haven't had a reply since"

I really want faster than my 75Mb but i do not want to go to a package that has yearly increases. Ofcom need to sort it out.

That other forum linked above mentions Aquiss resell City Fiber though, yet CityFiber isn't available here but Aquiss is so do they use multiple services, pretty sure only Openreach is available here.

They sell Openreach FTTP, and the backhaul is Entanet (which is now part of Cityfibre).

 

[Rainmaker]

yeah my UX shows cityfibre aswell, even though openreach van came to do the install, i didnt think too much of it i just assumed it was how the router picks it up.
cityfibre isnt even available on my road yet.

Openreach install all Openreach based FTTP. The line is Openreach to the exchange, handed off to the ISP and then onto whatever backhaul they use.

 

[Rainmaker]

It's not Cityfibre, Aquiss use an Entanet wholesale platform which Cityfibre bought.

Yup, as per my previous post.

 

[Rainmaker]

I did post this above but in their listing on Companies House, Aquiss only have one employee. That’s why the MD is answering the queries, he is the ONLY employee. He also makes the coffee for the whole office. Because he is the whole office.

It’s not a bad thing to be a growing business but if he gets sick response times could drop off a cliff. I am looking really hard at Aquiss because I manage a few connections for co-workers and currently they are mainly Zen, who appear to be going downhill at a canter. So Aquiss are a possibility, as are A&A.

It's a small family business, so I assume there's some 'ad-hoc' stuff going on. For 2023 there was 1 employee, for 2022 there were 0 employees; so I assume above and beyond Martin and the secretary respectively. I had a ticket reply the other day from one staff member, and a follow up from Martin later that day apologising for being in meetings and actually answering the query, so i don't know. Either way they've been going years and seem decent to me.

 

[Rainmaker]

just a heads up if you haven't seen the WhatsApp status, there are some planned maintenance this week for Tuesday, Wednesday and Thursday night while they do some updates.

Thanks, I hadn't actually seen that (I use Signal on the daily, and have WhatsApp for when really needed). Hopefully once this round of updates is complete they won't need doing for some time. I notice they've had quite a few of them lately, London, Cardiff etc. I wonder what routers they actually use? I'll ask Martin next time we speak, if I remember. I imagine something closer to a Juniper MX304 than a Mikrotik hEx, but who knows? I've seen some resellers/carriers using OpenBSD on x86 and all sorts of weird and wonderful things in between. If it works, it works!

 

[Rainmaker]

yeah would be good to know, i have seen they are quite frequent lately compared to never having an outage for years, and its during the time of the evening that i would be using internet.

Quite. We'll soon find out, but I hope it's not too disruptive. I'm assuming the quoted times are worst-case scenario in case something goes bork. In theory a core or edge router upgrade shouldn't take more than a few minutes, at least of actual downtime. Fingers crossed...

 

[Rainmaker]

Indeed, but they (Martin) does seem to start on the beginning time for the last couple of maintenance periods which is not ideal for domestic users. I've got VM as secondary/failover WAN so I barely notice other than notifications, but if it was my only WAN I wouldn't be happy especially for the quality service price.

I am thinking of posting about this on their forum to ask if Martin can try some Pro Plus or sleeping in, but also so I could be the second forum thread

I'm a night owl (practically nocturnal tbh). Maybe I should offer to do it for him at 3am-5am.

 

[Rainmaker]

Absolutely - it’s effectively how I operate myself. And I’m not knocking the guy or Aquiss, it’s literally when folks say “my support ticket was answered by the MD”. Well, yes, because he answers all the support tickets.

As I said earlier, to put it another way - I no longer have to navigate a seven-layer-deep menu, wait on hold for an hour and a half to speak to someone offshore, so they can tell me to reboot the modem (again) and then randomly cut me off 15 mins in after having me re-do the most basic of troubleshooting (after I told them I already did it). With an outfit like Aquiss, it's literally just answered within 2 rings and 'Afternoon... The PPPoE session isn't authenticating' (or whatever), and someone who knows what they're on about says 'Oh try this... Fixed? Have a good one!' and you're off the phone in 2 minutes (well, not counting random chat about networks/ISPs/OcUK).

 

[Rainmaker]

absolutely you are probably right, its just not a good look when i'm telling my mates about this great isp they have never heard of, and i'm the only one who cant get online for a period of time .

It's OK, if they're on Virgin they won't be online to notice Aquiss' router rebooting anyway.

 

[Rainmaker]

Martin has posted a status update on the WhatsApp channel to clarify that the upgrades are not for Aquiss directly, but from upstream. I don't want to copy the whole thing here as it's not for me to do, but he has already fed back that the start time would ideally be later to avoid potential disruption for customers. He also clarified that the upgrades and outages are infrequent bordering on seldom across all services (I'm paraphrasing here), and that Aquiss prefer to publish them where most ISPs don't bother, as he believes in transparency with customers. Sounds absolutely fair enough to me, and I'd rather know than guess! Another +1 for Aquiss tbh.

 

[Rainmaker]

Do they use what used to be Entanet (https://www.peeringdb.com/net/155) for backhaul across the UK and at what point do they handover from Openreach?

The prices/reviews look decent, but I do have some concerns about Entanet and their network resiliency

You'd be best placed dropping them a mail and asking; but yes it's Enta/CityFibre backhaul, and afaik (which isn't much) it hands off from Openreach at the exchange. As I said though, only they can actually answer you.

Anyone know if you can add the extra static IPs (/29) at any time during your contract or do you have to do it at the start?

Rather the opposite. You can't order from the outset, but once you're live you can go into your account (eBilling) and select it as an 'upgrade' for £5 a month. IPv6 is a free upgrade option.

 

[Rainmaker]

Ah, good to know, thanks! Do you know if you can upgrade/downgrade it whenever you want or is it fixed for the rest of your contract once you upgrade it?

I have the option to remove the /29 and get a refund for unused time, so I imagine you can. I've only just activated it though, so I can't say how permanent. Again, Aquiss are the ones who can answer this definitively but that's what I know.

 

[Rainmaker]

No, my /32 was on 78.something and my /29 block was on 80.something

There's no notification as such, you have to check in the Aquiss billing portal to see the new allocation.

Mine were sent over this morning, though it's been a busy one so I haven't had time to do anything about it (yet) except reboot the ONT. I didn't get a /29 routed over a /32. I got a straight /29, losing the 1st to gateway and the last to broadcast. My static WAN IP by default is the 2nd IP in the range, leaving the rest for static assignment.

 

[Rainmaker]

Can i ask how you've setup your config ?

The config you've posted is for DSL, for Aquiss/FTTP you want it like this:

/etc/config/network

config interface 'loopback'
        option device 'lo'
        option proto 'static'
        option ipaddr '127.0.0.1'
        option netmask '255.0.0.0'

config globals 'globals'
        option packet_steering '1'

config device
        option name 'br-lan'
        option type 'bridge'
        list ports 'eth1'

config interface 'lan'
        option device 'br-lan'
        option proto 'static'
        option ipaddr '10.100.0.1'
        option netmask '255.255.255.0'
        option ip6assign '64' # I assigned a /64 to LAN because it's massive already and is a 'normal' size IPv6 network
        option force_link '0'

config interface 'wan'
        option proto 'pppoe'
        option device 'eth0'
        option username '[email protected]'
        option password '123password'
        option ipv6 'auto'
        option peerdns '0'  # I disabled Aquiss' DNS in favour of a a private one (for the router itself, not LAN)
        list dns '76.76.2.41'
        list dns '76.76.10.41'
        option delegate '0'

config interface 'wan_29' # My static IPv4 /29 subnet minus the network and broadcast addresses, and the 1st address which is allocated to PPPoE
        option proto 'static'
        option device '@wan'
        list ipaddr '78.xxx.xxx.98/29'
        list ipaddr '78.xxx.xxx.99/29'
        list ipaddr '78.xxx.xxx.100/29'
        list ipaddr '78.xxx.xxx.101/29'
        list ipaddr '78.xxx.xxx.102/29'

/etc/config/dhcp

config dnsmasq
        option localise_queries '1'
        option rebind_protection '1'
        option local '/lan/'
        option domain 'lan'
        option expandhosts '1'
        option cachesize '10000'
        option authoritative '1'
        option readethers '1'
        option leasefile '/tmp/dhcp.leases'
        option resolvfile '/tmp/resolv.conf.d/resolv.conf.auto'
        option localservice '1'
        option ednspacket_max '1232'
        list interface 'lan'
        option dnsforwardmax '10000'
        option rebind_localhost '1'
        list rebind_domain 'plex.tv'
        list rebind_domain 'my.domain'

config dhcp 'lan'
        option interface 'lan'
        option start '10.100.0.200'
        option limit '50'
        option leasetime '12h'
        option dhcpv4 'server'
        list dhcp_option '6,10.100.0.50'
        option ra 'server'
        option ra_default '1'
        list dns '1004:1aa6:cd0f:6142:ad31:12cd:fe00:a123' # This is redacted/made up, real IPv6 LAN DNS server goes here
        list ra_flags 'other-config'

config odhcpd 'odhcpd'
        option maindhcp '0'
        option leasefile '/tmp/hosts/odhcpd'
        option leasetrigger '/usr/sbin/odhcpd-update'
        option loglevel '4'

config host
        option name 'core-switch'
        option dns '1'
        option mac '11:AA:23:BB:45:CD'
        option ip '10.100.0.2'

config host
        option name 'etc etc'

It's a lot faster to set up from the GUI, though!... Just set WAN to PPPoE, and IPv6 to automatic. That will generate a wan_6 alias interface (which doesn't show up in /etc/network) and then you can delegate the prefix and announce it accordingly on the LAN interface.

 

[Rainmaker]

I was affected by this too. Speeds were down to around 30Mbps to most destinations (but not all, I still had full speed via Switzerland), and packet loss kept piling up. Seems to be back to normal today touch wood! The spikes after lunch today are me restarting the PPPoE session (and then the router itself, to be safe).