Re: ISDN with BitSurfr
Alex (alex@livingston.com)
Tue, 22 Oct 1996 14:50:30 -0700
At 03:50 PM 10/22/96 EDT, you wrote:
>
> Well, I bought one of those TC-200-S6-1 cards (16650 UART, 460Kb max DTE) to
>use under Linux with my BitSurfr Pro. After a little fiddling I got my
>scripts to work again (the fact that the BitSurfr can't autobaud into or out
>of 230Kb makes it a hassle. Also, you have to "trick" the Linux kernel into
>doing 230Kb because it doesn't yet know about that speed. That bit is easy).
>
> All-in-all, I was able to increase my NetPipes (a raw TCP stream connection)
>throughput from 9.1KB/sec to 10KB/sec by moving to the higher DTE. This still
>isn't the 14KB/sec FTP times my friend with a pair of Ciscos reports over his
>ISDN line.
>
> Maybe the Cisco is doing compression.
> Maybe I have some stupid arguments to pppd.
> Maybe the BitSurfr isn't terribly good at saturating either link.
> Maybe the PortMaster isn't terribly good at saturating either link.
> Maybe that's near the theoretical limit of an ISDN connection after you
>factor in the TCP overhead.
> Maybe I have exposed inefficiencies in Linux's software.
> Maybe BellSouth's switch isn't really giving me 128Kb.
> Maybe "dd" isn't efficient at spewing/slurping data (but it's good for
>counting out bytes from /dev/zeros).
> Maybe my cat is chewing on grass right now.
>
> I don't know. I'm disappointed, but it was only $33 after shipping, so I'm
>not jumping off any bridges (not that there's anything high enough in Florida
>to jump off and kill yourself).
I remember that when we were testing the Bitsurfr Pro many months ago, we
found that it did not perform frame fragmentation very efficiently. Also,
it seemed to perform a lot of flow controlling on the data. Both of these
things cut down its overall throughput on two B channels. I know for a fact
that we can saturate the pipes--I've done the test myself with our in house
script.
alex
Alex Henthorn Livingston Enterprises
Senior Technical Product Manager 6920 Koll Center Parkway #220
Product Marketing Engineer Pleasanton, CA 94566
alex@livingston.com Voice 510-485-6787 Fax 510-244-1903