cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
2114
Views
5
Helpful
11
Replies

Switch Virtual Interface (SVI)

davidhuynh5
Level 1
Level 1

What is the purpose of an SVI and can I have two SVIs on the same vlan, but different switches trunked together?

SW#1:

inter vlan 150

ip address 192.168.0.1 255.255.255.0

SW#2:

inter vlan 150

ip address 192.168.0.2 255.255.255.0

Thanks in advance.

1 Accepted Solution

Accepted Solutions

David

Your follow up question was how a host would know which gateway to use. And the answer is that each host would be configured with a default gateway. Who ever configured the host would choose 192.168.0.1 or they would choose 192.168.0.2. And whatever gateway was configured will be the one that the host uses.

I would guess that what your CCIE was talking about was to systematically alter configuring the hosts so that half of them use 192.168.0.1 and the other half use 192.168.0.2. This would produce load balancing where half of the hosts use one switch while the other half use the other switch. This is essentially a manual solution to the question of how to load balance.

HTH

Rick

HTH

Rick

View solution in original post

11 Replies 11

Richard Burts
Hall of Fame
Hall of Fame

David

The purpose of an SVI is to be a layer 3 interface for processing a particular VLAN on the switch.

Yes it is quite possible to have interface vlan 150 on both switch 1 and switch 2, as you show, as long as each SVI has a unique address within the subnet.

HTH

Rick

HTH

Rick

Rick,

If a host with a single gateway is connected to VLAN 150, how will it know which SVI to route out?

An CCIE told me that it would be good for load balancing? Sorry, I'm just trying to get a grasp.

David,

Using HSRP as an example, a virtual ip address is assigned to your SVIs and these are used as the gateway for your hosts:

SW1#

int vlan 150

ip address 192.168.1.2 255.255.255.0

standby 150 ip 192.168.1.1

standby 150 preempt

int vlan 150

ip address 192.168.1.3 255.255.255.0

standby 150 ip 192.168.1.1

standby 150 preempt

The hosts use 192.168.1.1 as their gateway, and there is one switch that's a primary and the other is the standby. If one switch goes down, the other will answer for it.

In the case of GLBP, it can load balance links whereas HSRP doesn't (it can, but not without creating more ips and having hosts use different gateways).

HTH,

John

HTH, John *** Please rate all useful posts ***

John Blakley
VIP Alumni
VIP Alumni

David,

I completely misread your post, so if you received my response, please disregard. I thought you were asking if two SVIs on the SAME switch could exist for the same vlan with ip addresses in the same subnet. Rick is right, and what you have will work. In fact, that's the way it should be in the case of using HSRP.

Sorry for the confusion :)

John

HTH, John *** Please rate all useful posts ***

David:

"Rick,

If a host with a single gateway is connected to VLAN 150, how will it know which SVI to route out?"

You can implement a redundancy solution, where the hosts on a LAN will not point to any particular router's SVI IP address, but instead to a virtual IP address that both routers may have control over.

The most common solution is known as HSRP, in which both routers are placed in a redundnacy "standby" group, but only one will control the virtual IP address (VIP) at any given time. If that router fails, the other takes over.

This is why you do not point the hosts on the LAN to a particular SVI, but to the HSRP virtual IP address that they both share. In other words, the default gateway of the hosts will be the VIP.

In the case of load balancing, there is another first hop redundancy protocol known as GLBP, which allows all routers in the redundnacy group to receive user traffic and forward it simultaneously. This utilizes network bandwidth more efficiently since both routers can use their connections and forwarding capability at the same time, instead of one at a time, as is the case with HSRP.

HTH

Victor

Rick, John, and Victor.

I understand HSRP and I pointed out to the CCIE engineer. He said NO HSRP, but load balancing. I ask him to clarify, but his response was like a foreign language to me.

David,

The only load balancing protocol that also provides redundancy is GLBP.

John

HTH, John *** Please rate all useful posts ***

David:

I cannot speak for your CCIE. I wasnt there. I can only guess that if he mentioned "load balancing" in the scenario with which you present us, he was probably referring to GLBP.

Read my post.

From above post:

"In the case of load balancing, there is another first hop redundancy protocol known as GLBP, which allows all routers in the redundnacy group to receive user traffic and forward it simultaneously. This utilizes network bandwidth more efficiently since both routers can use their connections and forwarding capability at the same time, instead of one at a time, as is the case with HSRP."

Thanks

Victor

David

Your follow up question was how a host would know which gateway to use. And the answer is that each host would be configured with a default gateway. Who ever configured the host would choose 192.168.0.1 or they would choose 192.168.0.2. And whatever gateway was configured will be the one that the host uses.

I would guess that what your CCIE was talking about was to systematically alter configuring the hosts so that half of them use 192.168.0.1 and the other half use 192.168.0.2. This would produce load balancing where half of the hosts use one switch while the other half use the other switch. This is essentially a manual solution to the question of how to load balance.

HTH

Rick

HTH

Rick

David:

"I would guess that what your CCIE was talking about was to systematically alter configuring the hosts so that half of them use 192.168.0.1 and the other half use 192.168.0.2. This would produce load balancing where half of the hosts use one switch while the other half use the other switch. This is essentially a manual solution to the question of how to load balance."

If this is what the CCIE suggested, he should be defrocked. :-)

Yes, this would give you some primitive form of load sharing, but, for one thing, it would depend on who the heavy talkers are.

You also would have no redundancy for the LAN hosts.

Victor

I think you nailed it on the spot.

That would be the only reason I could think of, but wanted to see verify that I wasn't losing my mind.

Getting Started

Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community:

Innovations in Cisco Full Stack Observability - A new webinar from Cisco