In reply to a post by eckiedoo:

Interesting in that it looks almost a simple aggregation of the two lines, as opposed to some more complex arithmetic, eg a Root 2 (1.414) improvement above the simple aggregation, ie around 7 Mbps.

On the other hand, 5 Mbps looks of the right order for 5 km.

Yeah. We have a third line that's kept as a backup and it's rather ironic that the downstream throughput on that was roughly the same as the old connection. Better sometimes. It could never match the upstream throughput of the bonded lines though and for us that was very important. We work as part of a trans-Atlantic software development team so we push almost as much data to the US as we pull.

The two bonded lines were both pretty poor considering. Just luck of the draw I suppose but we never could get BT to do anything about it apart from when one went seriously bad. Otherwise it was the usual story from BT :-/

What we have now (for whatever it's worth) is a 100Mb fibre bearer to our little office park. That's then distributed around four companies. I think we all took a symmetrical 10Mb/s service in the end. Costs were £3k each for install and £220 pcm rental. Cheaper than we'd thought but it took a [censored] long time to get it installed. Organising all four companies was like herding cats and then we got caught up in Openreach's current woes so it took nearly five months instead of three months to go live.

Edited by Andrue (Sun 14-Jul-13 10:08:15)