I am attempting to run a full bitcoin node on premise. While I have a fairly large pipe, I am concerned that my ISP will not be tolerant of the added utilization that a node running full bore can consume. The last few days, it has averaged 15-20 GB/day in uploads. At the very least, I would like to know I have the ability to cap it's network utilization. I'm currently experimenting with using the linux traffic control script referenced below to cap bandwidth.
What I have noticed while using this is that it seems to create a QoS problem on my client WITHIN the bitcoin protocol. Assuming I understand the protocol correctly, what I'm seeing makes me think the client does not prioritize "core" network functions above torrenting historical blockchain data to other nodes. For example, I'm seeing high latency figures from bitnodes "getaddr" calls. I can mitigate this to some extent by restricting the number of connections, but it seems to me like the protocol could benefit from applying some algorithm to prioritize the core network functions over archived block sharing?
Any clever ways to mitigate this?
Traffic Control script:
https://github.com/bitcoin/bitcoin/blob/master/contrib/qos/tc.sh
