Multilink PPP connection

Multilink PPP connection

Post by Wisema » Mon, 23 Feb 1998 04:00:00



Hello,

I've been toying for a while with the mpd multilink ppp daemon version
1.0b4 (i've also tried with the pre-released 1.0b5) and seem to have
trouble getting any significant performance receiving data using 2 USR
sportster x2 links connecting to a US Robotics TotalControl System which
supposedly has Multilink PPP support. Both links dial, connect, and
transfer data, however i get WORSE performance with 2  x2 links than
with one. Sending, however is significantly better in performance than
with one x2 link.

As an example, receiving a 100k approx. binary executable through FTP
reports around a horrible 500 bytes to 1/2Kbytes/sec transfer rate.
Sending, however ranges from 5-9Kbytes/sec+ transfer rate. BTW yes, i've
tested with and without compression.

In addition to this, throughout the connection at random times (but
usually while receiving something), one or both modems exhibit an error
message and immediately proceed by one link* up and attempting to
regain connection:

mpd: [modem2] NEW FRAME ERRS: FCS 1 ADDR 0 RUNT 0 OVFL 0
mpd: [modem1] NEW FRAME ERRS: FCS 1 ADDR 0 RUNT 0 OVFL 0
mpd: [modem2] carrier detect (CD) signal lost
mpd: [modem2] Closing physical layer in state UP
mpd: [modem2] got msg MSG_DOWN
mpd: [modem2] LCP: Down event
mpd: [modem2] LCP: state change Opened --> Starting
mpd: [modem2] LCP: phase shift NETWORK --> DEAD
mpd: [foo] up: 1 link, total bandwidth 48000 bps (both links average
about 45-48k so settings are relative)

Not sure whether its something wrong with mpd or something server-side
related, but from what I see so far it could either be:

1. Local side can split packets up fine for sending, but server-side
can't split receiving packets up to send back to local side resulting in
packet loss, etc. which may also result in the modem errors/hangups.

2. Local side splits packets up fine for sending, server-side splits
receiving packets up fine to send back, but local side incorrectly
RECOMBINES incoming packets resulting in packet loss, etc. which may
have also resulted in the modem errors/hangups.

3. Incompatible MPPP standards between the daemon and the Total Control
system resulting in successful connection but unsuccessful performance.

Therefore, my question is whether anyone has successfully established
multilink ppp connections (experiencing similar problems or not) with
mpd and a USR total control system, if so, how? Also, whether the
problem resides in one of the 3 possible problems listed above or an
overlooked issue?

Any help would be appreciated.


 
 
 

1. Linux Multilink PPP connection with 2 dial-up modems

Finding information on this subject seems to be difficult.  Is EQL the only
Linux support for "Multilink PPP"?

Is EQL compatible with an ISP running a US Robotics Total Control Box?  If
anyone is aware of some detailed documentation on configuring Linux for
Multilink PPP, please let me know.

More importantly, when an ISP says they support Multilink PPP, does this
mean EQL will work, or is there some other method for doing Multilink?

Thanks,

Mike

2. Gnu Linker probs with Linux. HELP

3. how to setup a multilink ppp connection ???

4. Installing Win/Linux shared swap on old swap partition

5. Multilink PPP connection help

6. how to connect new disk tray to SSA200

7. can't establish multilink ppp connection

8. SuSE 8.2 - possibly a waste of money

9. Multilink PPP with one fixed connection, and one PPPoE connection

10. RH5.1 PPP connection: PPP connection persists, telnet session times out

11. What is the correct way to take down a Multilink connection?

12. Ask ppp-on and ppp-on-dialer for ppp connection

13. Multilink PPP