"I was assuming the local loop would be the bottleneck for upstream. There is only 2mb upstream capacity in the local loop, no?"
You're absolutely correct, except the data rate may vary with vendor, but that's the idea.
The bottleneck effect, however, is exacerbated with higher numbers of users. That's why I suggested that a remedy would be to group fewer users on the local loop portion, never overburdening any isolated segment with too many users, and then take the smaller grouped traffic and map it over dedicated bandwidth for that group, or multiplex may small groups in the upstream direction from onto a very high speed pipe from the field node to the h-e.
In the latter, you are effectively taking many lower speed lines which are functioning perfectly well (without creating congestion) and mounting them onto an express line to the head end. In this way, however, every lower speed line is used to its optimum level of performance without being overburdened by too many users. |
