Cable Forum - View Single Post

graf_von_anonym · 03-04-2009, 23:24

I don't have the spreadsheet on this computer, I'll have to dig it out, but the initial trial actually increased the amount that you could download in a day assuming you could saturate your connection 24 hours a day. I'll run the numbers on the new trial this weekend.

One thing that struck me just there was that I have only just realised that Virgin are putting new trunk (or trunk equivalent) between the new hardware for 50Mb and the rest of the internet, as otherwise it would only be making the situation worse. I assume that since the UBR/CMTS model basically collates a whole series of connections into the big pipe that the next upgrade sequence will be more on those trunks. Obviously switching from 4 to 8 multiplexing doubles the amounts that can be threaded through, but I don't know how big those pipes are. Obviously their upstream is effecively trivial as a load, because, hell, the fibre has room for light in two directions. Indeed, I'm not even sure if that serves as a constraint at all on the network, or if it's the routing/bandwidth of each individual UBR that brings it in.

Any of you network types care to chime in?

03-04-2009, 23:24	#29
graf_von_anonym Inactive Join Date: Dec 2008 Location: Castell Anonym Services: BB:NTL XL, TV:L Posts: 141	Re: STM Trial times changed I don't have the spreadsheet on this computer, I'll have to dig it out, but the initial trial actually increased the amount that you could download in a day assuming you could saturate your connection 24 hours a day. I'll run the numbers on the new trial this weekend. One thing that struck me just there was that I have only just realised that Virgin are putting new trunk (or trunk equivalent) between the new hardware for 50Mb and the rest of the internet, as otherwise it would only be making the situation worse. I assume that since the UBR/CMTS model basically collates a whole series of connections into the big pipe that the next upgrade sequence will be more on those trunks. Obviously switching from 4 to 8 multiplexing doubles the amounts that can be threaded through, but I don't know how big those pipes are. Obviously their upstream is effecively trivial as a load, because, hell, the fibre has room for light in two directions. Indeed, I'm not even sure if that serves as a constraint at all on the network, or if it's the routing/bandwidth of each individual UBR that brings it in. Any of you network types care to chime in?