South Africa’s biggest forum. Discuss, discover, and connect with thousands of members.
Out of neccessity. Even if it were on offer locally, to get anywhere near 24Mbps your loop must be of high quality, in a good state of repair and < 2000m (or <20dB ds attn).
Correct, locally avaliable solutions do not bond at the ADSL/PPPoE layer, since Telkom do not support this (on their ESRs). At higher layers, they have no say in the matter.
Ok yeah valid point, but still =P Sure 24mbps for one line is hard-pressed, but 10 mbps is reasonable. The idea of bonding like 4-8 ADSL lines just seems a bit absurd. I guess if it works and people are willing to pay and Telkom doesn't crack down on it, then what the heck... ;/ Another problem is that many exchanges are heavily contended... when I read "40Mbps bonded" I just pictured in my mind the problems that so many exchanges are having with a lack of backhaul capacity to the ESR... so ONE client rapes the exchange with his bonded connection and everyone else that's subject to contention suffers bad latency and speed degradation. You can't deny that bonding completely destroys contention at the DSLAM level... So fortunately it's out of reach for most home users in terms of cost.
I dont see how bonding has any bearing on the matter.
How can't you see that bonding will have a bad effect in terms of contention? The basic principle of contention ratios assumes that not all customers will be maxing out their lines at any given time, so an allocation of say 4Mbps could be divided amongst 4 users on a large-scale (so like 400Mbps total capacity for 400 users) so that even if those users each have 4Mbps packages, that 400Mbps would be enough for the customers to still obtain 4Mbps on their lines because not all lines will be utilized fully at any given time so that quality of service by average remains good for all customers; but as soon as you bring in bonding, then when one customers starts a download which maxes out all the lines which form that bond, then it certainly disrupts the contention ratios especially if that exchange's capacity is low... I doubt Telkom takes bonding into account when they work out the number of clients to supply per exchange based on that exchanges total bandwidth capacity (ie: the contention ratio)... so clearly bonding is asking for trouble here if those margins are slim
Not its not. Why do assume that the customer with 4 (bonded) lines will use more than the 4 individual ones?
