First impressions (based on half a day's results) are that your connection probably suffers from over-utilisation of VM's local network. I base this on the very high latency that magically improves at 12:30am - a classic symptom. The hourly spikes are normally where a connection is on speed test by VM, set up by their agents, and that's a separate issue.
Over-utilisation is equivalent to excessive contention on Openreach, although the details are subtly different. Basically, over-utilisation is where VM have insufficient capacity at peak times on the analogue radio frequency network between your hub and the CMTS that sits at the head of the local distribution network. Whilst there's various possible underlying causes, usually the latency is caused by packets being queued at the CMTS because traffic exceeds the real time processing capacity. The roots of these problems are protracted over-selling relative to available capacity. When VM recognise a problem they do try and stop selling, but these processes don't seem to happen early enough, added to which VM's tests for over-utilisation are very undemanding, hence the problem.
If my suspicion is correct, there's nothing you can do to improve matters, and this is unlikely to be an easily resolved noise or settings problem. In some areas VM do undertake work to rejig the local networks to balance loads and eliminate over-utilisation, and they do spend millions each year on capacity improvements. But sometimes/often that's judged uneconomic, or simply a lower priority than other investment options, and if your area isn't up there on the priority list, nothing will be done and the problem can persist for years. VM's track record on dealing with over-utilisation is very poor and often there's denial, unhelpful statements that "across a 24 hour period there's no problem", but even if there is a fault reference and a "fix date" there's no way of knowing if that fix date is actually backed by a real plan of action and programme of works. Quite often it seems not, and as the fix date approaches it is simply moved a month or two ahead.
This far I'm ahead of myself, we don't know if this is a utilisation problem, a few more days of the BQM will confirm. I am assuming that VM have repeatedly checked the hub from their side and assured you that all is well. Staff here operate under a strict policy that they must tow the company line, and that line says that unless over-utilisation has been formally diagnosed by the Networks team, then they must state that no problem exists. If my spidey-senses are correct then your options are limited:
1) Sit it out, and hope that VM do carry out improvement works. There's little or nothing you can do to force VM to upgrade the network, nor to be honest about the outlook. One or two customers have managed to drive progress on this front themselves, but typically we're talking seven months of brutal persistence, and interaction with VM every two or three days across that time. Even if you're quite remarkably focused, that's a level of commitment that will be a drag on your wellbeing.
2) Get yourself a new ISP. If you're in a fixed term contract you'll probably have to use the VM complaints process (and almost certainly escalate for arbitration at CISAS ) to be released from contract without penalty. If you need to do this, the grounds of your complaint is the poor performance, and your request for release from contract without penalty is twofold: First the Consumer Rights Act 2015 that requires any consumer service to be provided with "reasonable skill and care", and second, the Ofcom Fairness Commitments, that state "Customers’ services work as promised, reliably over time. If things go wrong providers give a prompt response to fix problems and take appropriate action to help their customers, which may include providing compensation where relevant. If providers can’t fix problems with core services they have promised to deliver within a reasonable period, customers can walk away from their contract with no penalty".
Let's see what the forum staff have to say, and the BQM shows over a couple of days. I could be (and occasionally have been) wrong about these matters, and the company may identify and fix a fault in short order. On the other hand, I'm right a lot more often than I'm wrong on these matters, and the persistence of the problem and ongoing failure of VM to fix it so far would suggest that the problem is one they choose to overlook, which is often the case with over-utilisation.
