seatrotter
Distinguished
As I recall, there are already network devices that are very capable of managing the bandwidth across a dynamic range of users. Say for the last minute, the device determined there were 50 users/ips (based on the upload stream packets). Based on that, it calculates BOTH the guaranteed bandwidth and burst/maximum/capped bandwidth for each ip. It does so every minute or n minutes.
So say the total limit is 50mbps. In the last minute, there were 50 active ips. The guaranteed bandwidth then is roughly 1mbps. The burst/maximum/capped bandwidth could either be the whole 50mbps or rule-/equation-based that could result to, say, 10mbps.
Of course, these devices are placed at, or near, the last node (where users/last mile connections are connected). Not only are they probably cheaper than DPIs and more automated than ISPs blindly throttling, but also it won't unnecessarily throttle users in areas where there are few active/heavy users.
So say the total limit is 50mbps. In the last minute, there were 50 active ips. The guaranteed bandwidth then is roughly 1mbps. The burst/maximum/capped bandwidth could either be the whole 50mbps or rule-/equation-based that could result to, say, 10mbps.
Of course, these devices are placed at, or near, the last node (where users/last mile connections are connected). Not only are they probably cheaper than DPIs and more automated than ISPs blindly throttling, but also it won't unnecessarily throttle users in areas where there are few active/heavy users.