This can be useful for measuring raw throughput: https://speedtest.telstra.com/
IIRC, iiNet and Internode are both owned by TPG these days. If you can only get 25/5 Mbs, that probably means you’re too far from the node. Given that you know there are line quality problems, the actual throughput will be much lower. Internet data travels in packets, comprising data and packet overheads such as destination address; packet sequence number; and a packet integrity checksum among other things.
When you have line errors, packets still arrive, but they’re corrupted. The checksum data check will detect this, and the protocol comms will cause the packet to be re-attempted. Hence, transferring a 5 MB unit of data over a bad line might require sending a much greater volume of data. If there are too many errors, the transfer will “time-out”.
In my experience, a better test than the Telstra one is to run a timed file upload of between 20Mb and 100Mb ready to upload to dropbox, and time it. This will give real world performance.
Does it really say “best endeavours” on the service offer? AFAIK, there have been court decisions on precisely what that means…e.g., I understand one judge concluded it means “not second best”. It is certainly well above and beyond “reasonable endeavours”, or “reasonable commercial endeavours”. It doesn’t mean “We’ll give it a try until we get sick of trying, then you’re on your own”.
Fixed wireless performance varies according to distance; interference; tower load. From what you’re saying, it sounds very much as if microwave is the weak link.
There really is no substitute for fibre. Nothing comes close.
The Abbott government debauched the once in a century infrastructure upgrade with its convoluted technology soup “stragedy”. It turned an long life investment into a short life expense, saving only a paltry amount of money in the short term, which will cost much more to fix later.
