From: asadovnikov (asadovnikov@comcast.net)
Date: Sat Jul 12 2003 - 15:09:53 GMT-3
My previous recommendation stands. Only thing it would not do is in option
#2 the upstream traffic will not be load balanced, which you can fix by
having 2 HSRP default gateways with 2 separate DHCP scopes. I would however
advice against it - it are not worse your efforts, especially if it is not
obvious what I am talking about.
Best regards,
Alexei
P.S. It sound like real-life design question for me. If it is a lab of
some sort please send more details.
-----Original Message-----
From: nobody@groupstudy.com [mailto:nobody@groupstudy.com]On Behalf Of
Emad
Sent: Saturday, July 12, 2003 7:17 AM
To: 'Larry Letterman'; boby2kusa@hotmail.com; 'Larson, Chris';
ccielab@groupstudy.com
Subject: RE: Switching: OT network design
Dear all,
What about if I have only one vlan on every edge switch instead of two
vlans!! How can I achieve load balancing between the edge switch and
these two core switches!!
-----Original Message-----
From: Larry Letterman [mailto:lletterm@cisco.com]
Sent: Friday, July 11, 2003 11:44 PM
To: 'Emad '; boby2kusa@hotmail.com; 'Larson, Chris';
ccielab@groupstudy.com
Subject: RE: Switching: OT network design
Both vlans are allowed on each trunk..but each trunk will be a
Root port only for one of the vlans ..based on the root core switch and
The root secondary core switch for that vlan..so in effect the trunks
will only
Carry the vlan data for the vlan that is going to root switch for that
vlan..
Larry Letterman
Cisco Systems
-----Original Message-----
From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of
Emad
Sent: Friday, July 11, 2003 1:37 PM
To: boby2kusa@hotmail.com; 'Larson, Chris'; ccielab@groupstudy.com
Subject: RE: Switching: OT network design
I just have one question,
All of you stated to have only one vlan is forwarding on each trunk ,
will that be the default behavior when I configure a trunk on both links
to the two core switches carrying both vlans and each core is the root
only for one vlans of the two carried by each trunk , or u are talking
about trunk pruning but it will not turn the runed vlans to forward when
one link fails , plz correct me
-----Original Message-----
From: boby2kusa@hotmail.com [mailto:boby2kusa@hotmail.com]
Sent: Friday, July 11, 2003 11:30 PM
To: Larson, Chris; 'Emad '; ccielab@groupstudy.com
Subject: Re: Switching: OT network design
for example,
6509 ----isl/dot1q----4506
| |
| isl/dot1q
|-------isl/dot1q-----4506
Before anyone say that Cat4K does not support ISL, I am assuming that
the sup is Sup3 or Sup4 which both support ISL and dot1q
The 6509 is the root only one of this port going to 4506 will be
forwarding for both vlans, so in order to achive load balancing you have
to make one port going to one 4506 forwarding for that vlan and blockign
for the other vlan. The same should be done for the other trunk but the
opposite vlan. I believe you can accomplish by changing the root port
cost or root port priority for each vlan.
I do not think pruning a vlan on a trunk would make it forward that vlan
if the other link goes down since it's being pruned.
----- Original Message -----
From: "Larson, Chris" <CLarson@usaid.gov>
To: "'Emad '" <emad@zakq8.com>; "Larson, Chris" <CLarson@usaid.gov>;
<boby2kusa@hotmail.com>; <ccielab@groupstudy.com>
Sent: Friday, July 11, 2003 1:21 PM
Subject: RE: Switching: OT network design
> Yes both uplinks will be trunked, but one of the vlans will be
blocked.
ie.
> on the uplink to core 1 even would be blocked, but it coudl carry both
1
and
> 2 (failure of core 2). So they will all be trunked.
>
> > -----Original Message-----
> > From: Emad [SMTP:emad@zakq8.com]
> > Sent: Friday, July 11, 2003 3:56 PM
> > To: 'Larson, Chris'; boby2kusa@hotmail.com; ccielab@groupstudy.com
> > Subject: RE: Switching: OT network design
> > Importance: High
> >
> > Larson
> > U are right in this point and I already know it , but I don't know
how
> > we will treat the two uplinks of the 4000 switch to the two core
> > switches , will each link carry only one vlan or will be trunk?
Because
> > I need redundancy also , if one link or one core switch failed , I
need
> > the another link to carry the traffic of both vlans to the another
core
> > switch
> >
> > Regards
> >
> > -----Original Message-----
> > From: Larson, Chris [mailto:CLarson@usaid.gov]
> > Sent: Friday, July 11, 2003 10:41 PM
> > To: 'boby2kusa@hotmail.com'; Emad; ccielab@groupstudy.com
> > Subject: RE: Switching: OT network design
> >
> > If each access layer switch has 2 vlans and a 1 gig uplink to 2 core
> > switches then the simplest thing to do is to make 1 core switch root
for
> > 1
> > of the vlans on each access layer switch and the other core root for
the
> > other vlan on each access-layer switch. Etherchannel is not
appropriate
> > because etherchannel can only be done to a single switch.
> >
> > In the Cisco switching guides they use even and odd vlans as an
example.
> > So
> > for instance if each access switch has consecutive vlans ie.
> > AccessSwitch1 = Vlan1, Vlan2 AccessSwitch2 = Vlan2, Vlan 3
> >
> > Then the config on core 1 is along the lines of
> >
> > set spanning root vlan 1,3,5,7,9
> > set spanning root vlan 2,4,6,8 secondary
> >
> > And on Core 2
> >
> > set spanning root vlan 2,4,6,8
> > set spanning root vlan 1,3,5,7,9
> >
> >
> >
> > Then you will "load balance" across both uplinks from each closet
switch
> > to
> > the core.
> >
> >
> >
> >
> >
> >
> >
> > > -----Original Message-----
> > > From: boby2kusa@hotmail.com [SMTP:boby2kusa@hotmail.com]
> > > Sent: Friday, July 11, 2003 3:22 PM
> > > To: Emad ; ccielab@groupstudy.com
> > > Subject: Re: Switching: OT network design
> > >
> > > ----- Original Message -----
> > > From: "Emad " <emad@zakq8.com>
> > > To: <ccielab@groupstudy.com>
> > > Sent: Friday, July 11, 2003 11:49 AM
> > > Subject: Switching: OT network design
> > >
> > >
> > > > Folks,
> > > >
> > > > I just wanna share an idea with you all ,
> > > >
> > > > I have 4506 switch in a network acting as the access layer and
> > connected
> > > > to two 6513 core switches with 2G uplink per each one,
> > > >
> > > > Each 4506 switch has 2 vlans and I want to have both uplinks to
the
> > core
> > > > switches working in load balance , how can we guarantee that:
> > > >
> > > > - By STP layer2 load balance? If yes , plz tell me how?
> > > Do you mean load balance the traffic thoughput between the 2 gig
> > uplinks
> > > or
> > > load balance the switches load?
> > > Load balancing between the 2 gig link can be accomplished by
> > channeling
> > > the
> > > uplinks. Load balancing the the switch's load (for a lack of a
better
> > > term)
> > > can be accomplished by having the one or the other as the bridge
to
> > the
> > > root
> > > (which should be the 6509). For example, 2 vlans, vlan 1 will be
> > > forwarding on 4506 A while blocking the vlan 2 and vice versa for
> > > the other
> > switch.
> > > This would be manipulating either the RP cost or the RP priority,
> > somebody
> > > will correct me if this is the wrong way to manipulate which
switch
> > should
> > > the vlan take on it's way to the root.
> > >
> > > > - By enabling routing protocol between the access layer and the
core
> > > > layer , but how?
> > > This would be load balancing on layer 3 and you would load balance
> > > according to the destination of the traffic, routers look at the
> > > routing
table
> > to
> > > forward the packet.
> > >
> > > >
> > > > I read the good paper of CISCO AVVID network infrastructure but
I
> > didn't
> > > > get it because most of scenarios are depending on one vlan and
> > > > redundancy between the two uplinks not load balancing and 2
vlans,
> > > >
> > > > Plz advice
> > > >
> > > > Thanx
> > > >
> > > >
> > > >
> >
This archive was generated by hypermail 2.1.4 : Wed Aug 06 2003 - 06:52:37 GMT-3