Hi Poe,

[Thus spake Poe Lim unto Sunny Leung:]

 SL> I wouldn't have a clue why this performance differential exists,
 SL> and AFAIK (using ATI6) connects with Netcomms do not utilise SREJ
 SL> which could have explained any performance gap.

 PL> I think Netcomms may use the larger block size for v.42; could also be
 PL> faster throughput at the other end too. 

That would explain it. I had a 31.2k/31.2k connect the other day to my
ISP. Boy was that good :-) Too bad it didn't shift up to 33.6k.

Sunny

... Audi alteram partem.
--- GEcho 1.11+