Ascend Archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: (ASCEND) mixing MLPPP channels w/ 6.0 software. (fwd)



At 03:37 PM 3/20/98 -0500, Jim Howard wrote:
<snip, snip, snip>
>
>While I'd agree that MP bundling a full T1 with a 64k ISDN channel
>might be messy, the differences in a 64k HDLC and 33.6k modem link
>are probably too small to notice with only a few bundled anyway.

It will also depend highly on the implementation at each end and how they
handle the allocation of packets/fragments over the channels in the
bundle.

>As long as the devices on each end are not trying to do some sort of
>time division or bit-level multiplex,where half a byte goes one path 
>and half goes another (ie, not a Multilink PPP),

Actually that's not totally true. It is not bit-level multiplexing, but
you can fragment packets over MP. It may be done in such a way that the
fragment sizes are proportional to the speed of the link for example.

>and instead do something more intelegent like queue an entire packet 
>for one physical link, and round-robin which link to send with
>in a "next available" scheme, all should be fine even if different speeds.

That's how the MAX will handle it should a client call in with one
ISDN and one modem session from the same endpoint.

>I have not actually tried this with ISDN and modem links,
>however I have tested with two K56flex modems (at 42-44k) 
>plus two V.34 modems (at 26.4-28.8k) in an MP bundle 
>doing all sorts of file downloads, streaming video, etc.,
>and never had any problems... performance was about equal to 128k ISDN.
>(using a BSD/OS 3.0 box as a router, and not with Win95/98/NT, BTW)

Into a MAX? Or was this BSD to BSD?


Kevin


++ Ascend Users Mailing List ++
To unsubscribe:	send unsubscribe to ascend-users-request@bungi.com
To get FAQ'd:	<http://www.nealis.net/ascend/faq>


References: