Not dumb at all, one of the tricky things in redundant links AND
redundant routers is default routing, redundancy and load sharing are
NOT good bedfellows.  With two providers and two routers I generally
configure full BGP routing, IBGP between the two routers and HSRP.  The
firewall behind the router defaults to the active router and that router
either forwards out it's serial or sends the packtet to the other BGP
speaking router based on it's routing table.

  The question asked on this post had only one router with two serial
connections, which is why I suggested dual default routes, BGP is
unecessary overengineering.

  

Gregg Malcolm wrote:
> 
> Dave,
> 
> I have a dumb question regarding multiple defaults.  Lets say that you had
a
> multihomed BGP config connected to 2 different providers.  Lets say that
you
> had 2 routers below the firewalls sourcing the default.  Then take a look
at
> the routing tables below these 2 routers. Wouldn't nearly every routing
> proto (other than RIP assuming the hop counts were the same) only list 1
> default? Wouldn't it be true that outbound traffic patterns would be based
> upon metrics from the routers sourcing the default?  If this is true, then
> it's not really load balancing.  I can think of scenarios were nearly all
> outbound traffic would be destined for only 1 of the 2 links.  I'm sure I'm
> missing something dumb, but figured it was worth asking anyway.
> 
> Gregg
> ""MADMAN""  wrote in message
> [EMAIL PROTECTED]">news:[EMAIL PROTECTED]...
> > First problem, BGP doesn't load share but with IOS you can source an
> > interface like a loopback, see BGP and loadsharing.
> >
> >   If you have two parallel paths to a single provider why are you doing
> > BGP???  Since you choose BGP I'll assume this is an Internet connection,
> > set up two default routes, ip cef global command and the configs you
> > have sent and you will have symetrical outgoing loadsharing.
> >
> >   Dave
> >
> > Alejandro Acosta wrote:
> > >
> > > Hi All,
> > >   This is my first message in the list.
> > >   I am running a BGP session with a customer. It has 2 serial links
with
> us
> > > (Each link of 2 Mbps). The customer and me have selected per-packet
> sharing
> > > in order to balanced the link.
> > >   In this moment, the traffic that comes from the customer is very
> simetric
> > > in both links, however, the traffic that is sent to the customer from
us
> is
> > > not simetric. As far as I know (if I am not wrong), if we are using
load
> > > balacing per-packet, the incoming and outgoing traffic should be very
> very
> > > similar, right?. Why only the incoming traffic is simetric in this
> moment.
> > >
> > > This is the configuration for both interfaces in my router:
> > >
> > > interface Serial2/0
> > >  description Link 1
> > >  bandwidth 2048
> > >  ip address xxx.xxx.xxx.xxx xxx.xxx.xxx.xxx
> > >  no ip directed-broadcast
> > >  ip load-sharing per-packet
> > >  no ip mroute-cache
> > >  load-interval 30
> > >  no cdp enable
> > >  hold-queue 1024 out
> > > !
> > >
> > > interface Serial2/4
> > >  description Link number 2
> > >  bandwidth 2048
> > >  ip address xxx.xxx.xxx.xxx xxx.xxx.xxx.xxx
> > >  no ip directed-broadcast
> > >  ip load-sharing per-packet
> > >  no ip mroute-cache
> > >  load-interval 30
> > >  no fair-queue
> > >  no cdp enable
> > >
> > > Any ideas?
> > >
> > > Thanks
> > >
> > > Alejandro Acosta
> > >
> > > P.D. I am using IOS 12.0(7)T
> > --
> > David Madland
> > Sr. Network Engineer
> > CCIE# 2016
> > Qwest Communications Int. Inc.
> > [EMAIL PROTECTED]
> > 612-664-3367
> >
> > "Emotion should reflect reason not guide it"
-- 
David Madland
Sr. Network Engineer
CCIE# 2016
Qwest Communications Int. Inc.
[EMAIL PROTECTED]
612-664-3367

"Emotion should reflect reason not guide it"




Message Posted at:
http://www.groupstudy.com/form/read.php?f=7&i=28993&t=28960
--------------------------------------------------
FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html
Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]

Reply via email to