From: Mas Kato (loomis_towcar@xxxxxxxxxxxxxx)
Date: Sat Apr 20 2002 - 20:31:09 GMT-3
[demime could not interpret encoding binary - treating as plain text]
Ya,
Given the exact wording and no other information, I think a carefully worded qu
estion to the proctor would be in order. One could "equally distribute your tra
ffic among both channels" at either layer 2 (ppp fragmentation) or layer 3 (no
ip route-cache).
It's entirely possible that the proctor may not care. Fine, designer's choice t
hen, but at least you know for sure. But there may also be other factors that m
ight influence your choice--perhaps some other requirement for PPP encapsulatio
n or wording in the task (or another task) that might preclude turning route-ca
ching off.
IMHO, resolving issues like this are a significant part of testing our expertis
e. It's about getting to the heart of the task-at-hand, determining what they r
eally want, figuring out any dependencies and then implementing the optimal sol
ution--quickly.
Regards,
Mas
> "yakout esmat" <yesmat@iprimus.com.au> "Mas Kato" <loomis_towcar@speedracer.c
om>Cc: <ccielab@groupstudy.com>
> RE: ISDN PPP MultilinkDate: Fri, 19 Apr 2002 19:40:31 +1000
>
>Mas,
>
>Good point.
>
>In first scenario "dialer load-thresold", what you are saying is that
>traffic will be load balanced in any case whether per destination (default)
>or per packet. But the router will treat both B channels as two separate
>links.
>
>In second scenario "PPP Multilink", the router will treat both B channels as
>ONE channel, i.e. the concept of load balancing -as we know it in layer 3-
>doesn't exist.
>
>In a scenario, If you've been asked to bring up the second B channel and
>make sure to equally distribute your traffic among both channels, which
>option would you go for.
>
>I would say we have to use "PPP Multilink"
>
>Thanks for your feedback
>
>Ya
>
>
>
>-----Original Message-----
>From: nobody@groupstudy.com [mailto:nobody@groupstudy.com]On Behalf Of
>Mas Kato
>Sent: Friday, April 19, 2002 4:15 PM
>To: yesmat@iprimus.com.au
>Cc: ccielab@groupstudy.com
>Subject: RE: ISDN PPP Multilink
>
>
>[demime could not interpret encoding binary - treating as plain text]
>Ya,
>
>'dialer load-threshold' sets the threshold for bringing up additional links,
>but each remains a discrete link. Any load balancing of traffic across the
>links is subject to normal IOS route-caching rules (on=per destination,
>off=per packet). Load balancing at layer-3, if you will.
>
>'ppp multilink' provides options on how to handle ppp encapsulated traffic,
>including a means to mux traffic across the two (or more) physical links as
>if they were one "virtual" pipe (fragmentation). Load balancing at layer-2,
>if you will. It also provides a means to control the way the traffic is
>"striped" across the links (interleaving and fragment-delay), which can
>prove useful when you are trying to share bandwidth between traffic with
>significantly differing packet sizes and delay sensitivities (say FTP and
>voice).
>
>Regards,
>
>Mas Kato
>https://ecardfile.com/id/mkato
>
>P.S. You are so lucky to live in Sydney! Man I love it there. -Cheers
>
>> "yakout esmat" <yesmat@iprimus.com.au> "Groupstudy"
><ccielab@groupstudy.com> ISDN PPP MultilinkDate: Wed, 17 Apr 2002 21:56:27
>+1000
>>Reply-To: "yakout esmat" <yesmat@iprimus.com.au>
>>
>>Hi all,
>>
>>I have a concept problem with PPP Multilink in BRI Interface.
>>
>>Cisco Documentation states:
>>
>>1) Don't have to use "PPP Multilink" to bring up the second BRI link when
>>needed, only "dialer load-threshold <n>" command is enough to do that, and
>>in this case we have to use "bandwidth" command to specify the interface
>>bandwidth so that Load-threshold can cancalculate the load on the first
>link
>>before bringing up the second one.
>>
>>2) If we use "PPP Multilink" command, "dialer load-threshold 1 " command
>can
>>no longer bring up both links
>>
>>My findings contradict all that.
>>
>>No difference between configuring "PPP Multilink" or not, I can still bring
>>up the other link by only using "dialer load-threshold <n>"
>>
>>I must be missing the point, can any body share views on this one.
>>
>>Appreciate your help in advance.
>>
>>Cheers from beautiful Sydney
>>
>>Ya
This archive was generated by hypermail 2.1.4 : Thu Jun 13 2002 - 10:58:14 GMT-3